# Agent Build: Configure How an AI Agent Works Source: https://docs.sketricgen.ai/ai-workforce/agent-build Use SketricGen Agent Build to configure a main agent, sub-agents, instructions, models, tools, skills, knowledge, handoffs, and deployment. Agent Build is the internal configuration surface for one AI Workforce agent. It shows how input reaches the main agent, which sub-agents it can delegate to, and which tools and skills each agent can use. Use Agent Build to change behavior. Use [Agent Space](/ai-workforce/agent-space) to see the agent working across the Project, and use [Playground](/ai-workforce/playground) to run and test it. Blog Writer Agent Build with an Input node, content-planning main agent, research tools, and specialist SEO sub-agent ## Agent Build Modes The top navigation separates three jobs: * **Build:** configure the agent graph and internal behavior. * **Run:** test the draft or published agent in Playground. * **Deploy:** publish or connect the agent to supported execution surfaces. The status near the top shows whether you are viewing the **Draft** or published version. Draft changes do not automatically replace the published configuration. Switching from Agent Build to the Playground Run experience ## Build the Agent Graph An Agent Build can contain: * **Input:** the message or instruction entering from the Run experience or a deployed surface. * **Main agent:** owns the user request, chooses tools, delegates work, and returns the final response. * **Sub-agents:** specialists that receive bounded tasks from the main agent. * **Tools:** capabilities such as File Search, Web Search, app connectors, API requests, structured output, and custom MCP tools. * **Skills:** reusable task instructions and supporting resources attached to an agent. * **Sticky notes:** documentation and setup reminders placed directly on the canvas. * **Connections:** define which tools or sub-agents the main agent can call. Keep the main agent responsible for the overall outcome. Add a sub-agent when a responsibility is distinct enough to deserve its own instructions, model, tools, and trace. ## Configure an Agent Select an agent node to open the configuration panel. Blog Writer AI Workforce configuration panel with content strategy instructions, model, tools, and skills You can configure: * **Name and icon:** identify the agent in Agent Space, Playground, and traces. * **Instructions:** define identity, responsibilities, process, tool rules, escalation, and expected output. * **Generate with AI:** draft or improve instructions, which you should review before saving. * **Model:** choose a model whose capabilities match the work. * **Tools:** attach only the actions and data sources the agent needs. * **Skills:** upload, browse, or generate reusable task guidance. Save changes before leaving the editor. Test one behavior change at a time so you can understand its effect in the trace. ## Main Agents and Sub-Agents The main agent talks to the user. A sub-agent works behind the main agent and returns its result to the main agent. Sub-agents cannot contain their own sub-agents. Use them for focused work such as research, classification, extraction, content drafting, validation, or escalation. Instructions should tell the main agent: * What result it owns * Which tasks it may delegate * When it should use each tool * When it must request human input * What it should do when evidence is missing See [Multi-Agent Orchestration and Handoffs](/user-guides/orchestration-and-handoffs) for delegation and routing patterns. To emit schema-validated JSON, add the dedicated [Structured Output tool](/dev-guide/structured-output-tools). ## Draft, Test, and Publish Use this improvement loop: 1. Change one instruction, model, tool, skill, or connection. 2. Save the draft. 3. Open **Run** and use a realistic test prompt. 4. Inspect tool calls, sub-agent delegation, files, human-input requests, and the final response. 5. Open the trace when the result is unexpected. 6. Publish only after the draft behaves as intended. ## Capability Facts * **Product surface:** Agent Build * **Purpose:** define the internal behavior and capabilities of one agent * **Contains:** input, main agent, optional sub-agents, tools, skills, notes, and connections * **Main agent:** owns the conversation and final answer * **Sub-agent:** receives delegated work and does not talk directly to the user * **Draft behavior:** editable and testable without automatically replacing the published version * **Run surface:** [Playground](/ai-workforce/playground) * **Operational surface:** [Agent Space](/ai-workforce/agent-space) ## Troubleshooting ### The agent does not use an attached tool Confirm that the tool is connected to the correct agent and that the instructions explain when to use it. Test with a prompt that clearly requires the tool, then inspect the trace. ### A sub-agent is never called Confirm that the sub-agent is connected to the main agent and has a distinct role. Tell the main agent which tasks should be delegated instead of duplicating the same responsibility in both agents. ### The Run result does not include the latest changes Confirm whether Playground is testing the Draft or published version. Save the Agent Build and start a new test conversation when you need a clean comparison. # Agent Space: Your Live AI Workplace Source: https://docs.sketricgen.ai/ai-workforce/agent-space Use SketricGen Agent Space to see AI agents, live work, schedules, teammates, quick conversations, and requests that need human attention. Agent Space is the live operational view for the selected Project. It shows which agents exist, what they are doing, who is working with them, and whether any run needs human attention. Agent Space is not the workflow editor. To change an agent's instructions, model, tools, skills, or sub-agents, open [Agent Build](/ai-workforce/agent-build). Filtering the Agent Space by operational status and opening an agent's live details This short tour shows the operational sequence: view every Project agent, focus on a status, then open one agent for live details and actions. ## Understand the Agent Space Each agent appears as a movable card on the Project floor. Your layout is saved for you, so teammates can arrange the same Project differently without changing the agents themselves. Use the filters at the top to focus on operational state: | Filter | Meaning | | ------------------- | ------------------------------------------------------- | | **All** | Every agent in the selected Project | | **Needs attention** | A run is paused and waiting for human input or approval | | **Working** | The agent has one or more active runs | | **Scheduled** | The agent has a deployed schedule or active automation | | **Sleeping** | The agent has no active work and no deployed schedule | An agent may display **Runtime unavailable** when live runtime status cannot be reached. That does not automatically mean the saved Agent Build was deleted. Agent Space overview followed by an agent details panel ## Open Agent Details Select an agent to open its operational details. Agent details with session, Agent Build, team, tools, and activity summary The details panel can show: * Current operational status * Number of sub-agents, skills, and tools * People currently working with the agent * Active runs * Requests waiting for human input * The agent's internal sub-agent team * Last activity * Deployed schedules and automations From this panel, select **Start Session** to open the Run experience or **Manage Agent Build** to change the agent's internal configuration. ## Work With Agents and Teammates Agent Space combines agent activity with human presence: * A teammate avatar means that person is viewing or working in a conversation with the agent. * Quick message opens a lightweight conversation entry point from the Project floor. * Selecting a teammate shows which agents and live runs they are working with. * Presence is informational; it does not grant additional Project permissions. For shared conversations, typing indicators, and parallel sessions with the same agent, see [Multiplayer AI and Team Collaboration](/ai-workforce/team-collaboration). ## Respond to "Needs attention" An agent enters **Needs attention** when a run pauses for a human-input request. Open the agent, select the waiting request, and continue in the associated Playground conversation. Human input can be a choice, clarification, confirmation, or approval. The run stays paused until an allowed user responds or cancels it. ## Create or Configure an Agent Select **Create new agent** to start a new Agent Build or choose a template. Creating an agent adds it to the selected Project and Agent Space. Use **Manage Agent Build** when you need to change: * Main-agent or sub-agent instructions * Model selection * Tools and connected apps * Skills * Knowledge access * Agent handoffs * Draft and published configuration ## Capability Facts * **Product surface:** Agent Space * **Purpose:** monitor and enter work across all agents in a Project * **Scope:** selected Teamspace and Project * **Shows:** agents, status, schedules, active runs, pending human input, and live collaborators * **Does not edit:** instructions, models, tools, skills, or the agent graph * **Layout behavior:** agent positions and viewport are saved per user * **Primary actions:** create an agent, start a session, quick message, open a request, or manage Agent Build * **Related surface:** [Agent Build](/ai-workforce/agent-build) ## Troubleshooting ### "The Agent Space could not be loaded" Refresh Agent Space. If the error continues, confirm that a Teamspace and Project are selected and that your membership has access to the Project. ### An agent does not appear under Scheduled The schedule must be deployed or the automation must be active. A draft schedule alone does not move the agent into the Scheduled filter. ### Live people or run status is unavailable The saved agents can still appear when live presence or runtime status is temporarily unavailable. Refresh the page before changing an Agent Build. # AI Workforce Overview Source: https://docs.sketricgen.ai/ai-workforce/overview Use SketricGen AI Workforce to configure agents, run collaborative work, use tools and files, request human input, and automate scheduled tasks. AI Workforce is SketricGen's internal agent system for research, marketing, content, sales, support, data, and operations. It combines a live Project workplace, visual agent configuration, conversations, files, schedules, tools, and human control. Use [Brand Agents](/brand-agents/overview) instead when the main user is a website visitor, prospect, customer, or public-channel audience. ## Core AI Workforce Surfaces ### Agent Space [Agent Space](/ai-workforce/agent-space) is the live workplace for the selected Project. It shows agents, operational status, teammates, active runs, schedules, and requests that need human attention. ### Agent Build [Agent Build](/ai-workforce/agent-build) defines how one agent works internally. Configure the main agent, sub-agents, instructions, models, tools, skills, knowledge, and handoffs. ### Playground [Playground](/ai-workforce/playground) is the Run experience. Start or continue conversations, attach files, observe work, respond to human-input requests, and inspect artifacts and traces. ### Multiplayer AI [Team Collaboration](/ai-workforce/team-collaboration) connects Agent Space and Playground. Teammates can see who is working with an agent, join a shared conversation, view same-conversation typing indicators, or use parallel sessions with the same Agent Build. ## What AI Workforce Agents Can Do Depending on the Agent Build and connected permissions, an agent can: * Plan and complete multi-step work * Search the web or an approved Knowledge Base * Use connected business apps * Call specialist sub-agents * Read, create, and edit supported files * Emit schema-validated data through a Structured Output tool * Pause for human input or approval * Continue longer-running tasks * Run manually or on a schedule * Share operational state with permitted Project members Capabilities are explicit. An agent can use only the models, tools, skills, knowledge, connections, files, and deployment surfaces available to its Project and Agent Build. ## Typical Uses * Market, competitor, and account research * Content planning and production * Prospecting and outreach preparation * Sales and meeting preparation * Internal support and knowledge assistance * Data analysis and report generation * Recurring operational checks * File transformation and structured extraction ## Start From Onboarding or a Template [Adaptive Onboarding](/getting-started/adaptive-onboarding) can recommend a relevant starting agent after learning your role, business, website, and goal. Templates provide a prebuilt Agent Build with instructions, agents, tools, setup notes, and test prompts. Review every connection and instruction before using a template with business data. ## Human Control and Transparency AI Workforce exposes activity through conversations, Agent Space status, traces, and generated files. A workflow can pause for human input before it proceeds. This visibility helps you distinguish: * What the user requested * Which agent handled each part * Which tools and files were used * What was generated * What failed or needs attention * Which human approved or continued the work when recorded ## Related Guides * [Agent Space](/ai-workforce/agent-space) * [Multiplayer AI and Team Collaboration](/ai-workforce/team-collaboration) * [Agent Build](/ai-workforce/agent-build) * [Playground](/ai-workforce/playground) * [Tools, Connectors, and Skills](/user-guides/tools) * [Runs, Schedules, Files, and Artifacts](/ai-workforce/runs-files-artifacts) * [Traces and Optimization](/user-guides/traces-and-optimizations) * [Multi-Agent Orchestration and Handoffs](/user-guides/orchestration-and-handoffs) # Playground: Run and Test an AI Agent Source: https://docs.sketricgen.ai/ai-workforce/playground Use SketricGen Playground to test agent drafts, manage conversations, upload files, review live work, provide human input, and inspect artifacts. Playground is the **Run** experience inside Agent Build. It lets you test the current agent, continue saved conversations, attach files, review generated artifacts, and respond when a run asks for human input. Playground is also where Agent Space sends you when you select **Start Session** or open a request under **Needs attention**. Playground showing a completed Blog Writer run, conversation history, prompt composer, and Project and agent workspaces ## Start a Test Conversation 1. Open an agent from Agent Space. 2. Select **Start Session**, or open Agent Build and select **Run**. 3. Confirm whether the header says you are testing the Draft or published workflow. 4. Select **New Chat** for a clean conversation. 5. Enter a realistic prompt that requires the behavior you want to verify. 6. Review the response, live activity, files, and trace before changing the Agent Build. Suggested prompts come from the Agent Build and help teammates test the intended use case consistently. ## Conversations The left panel lists conversations for the current agent. Conversations started from Playground, Agent Space, or a schedule can appear in the same signed-in history. Use a new conversation when you want to test without relying on earlier chat context. Reopen an existing conversation when the agent should continue work with that history. Conversation visibility can depend on Project membership and whether the conversation or automation is shared or private. A visible conversation does not grant permission to change the Agent Build. ## Prompt and File Input The composer supports: * Text instructions * File attachments * Voice input when available in the browser * File mentions using `@` * Suggested prompts defined for the agent Only attach files that the selected Project and agent are allowed to use. File access does not automatically give the agent access to unrelated connected apps. ## Workspace files and artifacts Playground gives you quick access to files used by the current conversation or run. The **Files** experience is the management surface and separates files by scope: Opening, collapsing, and restoring the Playground workspace panel * **Project files:** files shared with every agent and conversation in the selected Project. * **Agent files:** files shared across all conversations for the current agent. * **Conversation files:** files created or used for a specific conversation or run. They remain private to that conversation and are removed 30 days after the last write. In Files, navigate folders and review file size and modified time. When your Project permissions allow editing, use **New file**, **Upload**, **Save**, **Rename**, or **Delete**. New files and browser edits are text-only. Preview images and PDFs in the browser, and download file types that cannot be previewed. Text files larger than 1 MB stay view-only. If a file changes after you open it, Save and Rename report a conflict so you can reload before replacing newer work. Select a file in Playground or Files to preview or download it. Review generated reports, spreadsheets, images, documents, and other artifacts before sending or publishing them elsewhere. Expanded Playground workspace showing separate Project and agent file scopes ## Live Activity and Human Input During a run, Playground can show agent activity such as model work, tool calls, sub-agent delegation, and generated files. A longer task may continue after the first visible step. When an agent requests human input, the run pauses. The request may ask you to: * Choose one of several options * Provide missing information * Confirm a proposed action * Approve or reject a sensitive external step * Enter a custom response Respond in the same conversation. Agent Space shows the agent under **Needs attention** until the request is answered, cancelled, or superseded. ## Inspect and Improve If the result is wrong or incomplete: 1. Open the trace for the run. 2. Check which agent handled the request. 3. Review tools, sub-agent calls, errors, latency, and credit usage. 4. Check the files and evidence used by the agent. 5. Change one part of the Agent Build. 6. Repeat the same test in a new conversation. See [Traces and Optimization](/user-guides/traces-and-optimizations) for deeper debugging. ## Capability Facts * **Product surface:** Playground, shown as the Run tab in Agent Build * **Purpose:** test and work with an agent through conversations * **Inputs:** text, supported files, voice input, and human-input responses * **Outputs:** messages, tool activity, traces, and workspace artifacts * **Conversation sources:** Playground, Agent Space, and scheduled runs * **Human control:** a run can pause for a choice, clarification, confirmation, or approval * **File scopes:** Project, agent, and conversation/run workspace * **Configuration source:** current draft or published Agent Build, as shown in the header ## Troubleshooting ### A conversation appears to keep running after navigation Return to the same conversation. When detached-run support is available, Playground can reattach to the active run instead of starting a duplicate. ### A file is not available to another agent Confirm its workspace scope. An agent-workspace or conversation file is not automatically a Project file. ### The agent is waiting but no new response appears Check Agent Space for **Needs attention** and inspect the conversation for a human-input request. Refresh the conversation before starting another run. # Scheduled Runs, Files, and Artifacts Source: https://docs.sketricgen.ai/ai-workforce/runs-files-artifacts Use AI Workforce for scheduled automation, report automation, longer runs, planning, workspace files, generated artifacts, traces, and recurring internal work. AI Workforce runs can do more than return a single chat response. They can plan work, use tools, create files, produce artifacts you can inspect, and run recurring tasks on a schedule. Use scheduled runs for report automation, recurring research, operations checks, and other repeatable team workflows. AI Workforce playground for testing an internal workflow ## Longer Runs Some tasks finish in seconds. More complex AI Workforce tasks may take 10-30 minutes depending on: * The selected model * Number of agents involved * Tool and connector latency * File size and count * Whether the agent creates or edits artifacts Longer runs are useful for research, content, analysis, internal operations, and multi-step work. ## Scheduled Runs Schedules let AI Workforce agents automate recurring work by triggering a run at a configured time. Instead of asking the agent manually each time, you can turn repeatable internal work into a scheduled run. ### Create a schedule from the Input node Open the agent's **Agent Build**, then select the **Input** node on the Build canvas. Choose **Scheduled Trigger** to configure the message and cadence that will start each recurring run. Agent Build canvas with the Input node selected and Scheduled Trigger configuration visible 1. Open the agent from Agent Space and select **Manage Agent Build**. 2. On the **Build** canvas, select the **Input** node. 3. Choose **Scheduled Trigger**. 4. Enter the **User Message** the agent should receive whenever the schedule fires. 5. Set the recurrence to hourly, daily, weekly, or monthly, then configure the relevant minute, hour, day, and timezone settings. 6. Select **Deploy Trigger**. At every scheduled interval, SketricGen sends the configured User Message to the agent as its trigger text. The agent then handles it like a new instruction and runs with its current instructions, tools, knowledge, connections, and permissions. Write the User Message as a complete recurring instruction. Include the subject, time range, sources, output format, and destination when they matter. For example: `Review this week's support conversations, summarize the top five recurring issues, and save the report as a Markdown file.` Scheduled runs work well for: * Daily research or news briefs. * Weekly sales, support, or operations reports. * Automated report generation and recurring executive summaries. * Recurring content drafts. * Pipeline, CRM, or project-management checks. * Knowledge Base or document review tasks. * Any internal workflow where the agent should run the same instruction on a regular cadence. Treat a schedule like an automation contract. Give the agent a clear recurring instruction, connect the tools it needs, define the expected output, and review the first few runs in **Runs / Schedules** and traces before relying on it. ## Planning and Todos AI Workforce runs can maintain a todo/planning state during a run. This helps the agent break down a larger task and track progress. The plan is a working aid, not a guarantee. Always review the final output, generated files, and [traces](/user-guides/traces-and-optimizations) when accuracy matters. ## Workspace files AI Workforce uses three file scopes. Open **Files** in the workspace to browse the selected Project's files: * **Project files:** shared with every agent and conversation in the selected Project. * **Agent files:** reusable across all conversations for one agent. * **Conversation files:** private to one conversation and removed 30 days after the last write. The Files experience supports folder navigation, file size and modified-time details, creating text files, uploading files into the current folder, saving edits, renaming files, and deleting files when your Project permissions allow those actions. Supported text files can be edited in the browser, images and PDFs can be previewed, and other file types can be downloaded. Text files larger than 1 MB remain view-only in the browser. Save and rename actions check whether the file changed after it was opened. If another change is detected, the action reports a conflict so you can reload before replacing newer work. Files created for a run or schedule use the Conversation scope. Generated artifacts can include drafts, reports, summaries, transformed data, or other files created by tools and agents. ## Inspecting Artifacts When a run creates files, inspect them in the product before using them downstream. For important workflows, verify: * The file exists and opens. * The content matches the request. * The format is usable by the target system. * Any connector upload or sub-agent step used the intended file. ## Traces Traces show the run timeline: agent steps, sub-agent calls, tool calls, errors, latency, and credit usage. Use traces to understand why a run succeeded, failed, or took longer than expected. For tool configuration, see [AI Workforce Tools, Connectors, and Skills](/user-guides/tools). AI Workforce trace view with user messages, skills, agent steps, and tool calls # Multiplayer AI: Collaborate With Agents in Real Time Source: https://docs.sketricgen.ai/ai-workforce/team-collaboration Collaborate with teammates around the same SketricGen agent using live presence, shared conversations, typing indicators, and parallel sessions. SketricGen's multiplayer AI experience lets several members of the same Teamspace work with one agent at the same time. Agent Space shows where people are working, while Playground provides the conversations, live responses, files, and human-input controls used to complete the work. This is collaboration around a shared agent, not a single shared text box. Teammates can either join the same conversation or open separate sessions with the same Agent Build. ```mermaid theme={null} flowchart LR A[Teammate A] --> S[Shared agent] B[Teammate B] --> S C[Teammate C] --> S S --> C1[Shared conversation] S --> C2[Parallel conversation] C1 --> R1[Live agent run] C2 --> R2[Independent live run] R1 --> F[Agent Space presence] R2 --> F ``` ## Two Ways to Collaborate ### Join the same conversation Use a shared conversation when teammates need the same context and want to review one continuous thread. 1. One teammate opens an agent and starts a Playground session. 2. Agent Space shows that a person is working with the agent. 3. Another teammate selects the person's avatar or opens the agent details. 4. Select **Join conversation** to open that conversation. 5. Participants in the same conversation can see teammate presence and typing indicators. 6. The active agent response streams in the Playground session, and the saved conversation remains available to authorized Project members. Agent Space showing the Project's agents and the live people count Conversations in a shared Project are **Public** by default, so authorized teammates with access to the agent can view and join them. If a conversation should stay personal, click **Public** in Playground to make it **Private**. Only you can read the private conversation you started; other Project members cannot view or join it. ### Use parallel sessions Use separate conversations when teammates want the same agent to perform different tasks concurrently. Each conversation keeps its own chat context and run, while Agent Space groups the people and activity under the same agent. Examples include: * Two marketers asking the research agent about different campaigns * A support lead reviewing one escalation while an operator handles another * Several teammates generating separate reports from the same approved Agent Build * One teammate testing the draft while another reviews a published workflow Playground with conversation history, live chat, and the shared agent workspace ## What Teammates Can See Agent Space can show: * How many people are currently working with each agent * Teammate initials around the agent * Whether a teammate is viewing or typing * Which conversation the teammate has open * Live runs associated with that teammate and agent * A **Join conversation** action when a shared conversation is available The agent details panel also includes **People here**, **Working runs**, and **Needs input** counts. ## Presence and Typing Playground sends a lightweight presence update while the page is visible. Other authorized teammates can then see: * A presence bubble when someone else is viewing the agent * A teammate typing indicator when both people are in the same conversation * The teammate's display name or initials Presence is temporary. It clears automatically after a teammate leaves the page, changes agents, closes the browser, or stops sending live presence updates. Typing status is scoped to the same conversation. A teammate working in a different conversation can still appear under the agent's overall presence without appearing as a typing participant in your thread. ## Live Responses and Shared Context Every conversation has its own saved history and live run state: * Joining a shared conversation uses that conversation's existing context. * When an active run can be reattached, Playground resumes the live run instead of starting a duplicate. * Separate conversations can run in parallel without sharing their chat history. * Project files are read-only across Project agents. * Agent workspace files can be reused across conversations for that agent. * Conversation files remain scoped to the relevant conversation or run unless promoted to a broader workspace. For a shared conversation, coordinate before sending two new instructions at the same moment. Use separate conversations when teammates need independent concurrent tasks. ## Human Input and Handoffs When a run needs clarification or approval, Agent Space places the agent under **Needs attention**. An authorized teammate can open the waiting conversation and provide the requested input. Use the conversation and presence indicators to avoid duplicate approvals. The run continues after an allowed user answers, rejects, or cancels the request. ## Access and Privacy Multiplayer access follows the selected Teamspace, Project, and agent permissions: * A user must be an authorized Teamspace and Project member. * Opening Agent Space does not reveal conversations the user cannot access. * Presence records describe current activity; they do not grant additional permissions. * Joining a conversation uses the same agent authorization checks as starting a session. * Public website visitors do not appear as internal Teamspace collaborators. ## Troubleshooting ### A teammate does not appear in Agent Space Confirm that the teammate has the Playground open for the same agent and that both users are in the same Teamspace and Project. Wait a few seconds, then select **Refresh Agent Space**. ### The typing indicator does not appear Typing indicators only appear when both teammates are in the same conversation. Someone using the same agent in a separate conversation may still appear as present. ### Join conversation is not available The teammate may be viewing the agent without an active conversation, or your membership may not allow access to that conversation. Select **Open chat** to start your own session. ### Presence is temporarily unavailable Agent runs and saved conversations can continue even when the live presence service is unavailable. Refresh Agent Space before assuming the teammate left. ## Capability Facts * **Product surface:** Agent Space and Playground * **Presence scope:** agent, Teamspace, Project, and authorized user * **Typing scope:** participants viewing the same conversation * **Conversation modes:** shared thread or independent parallel sessions * **Live work:** agent responses, run state, files, and human-input requests * **Primary controls:** open chat, join conversation, start a separate session, or manage Agent Build * **Does not change:** agent instructions, models, tools, or permissions * **Related guides:** [Agent Space](/ai-workforce/agent-space), [Playground](/ai-workforce/playground), and [Agent Build](/ai-workforce/agent-build) # Brand Agent Connectors and Lead Capture Source: https://docs.sketricgen.ai/brand-agents/connectors-and-leads Add safe customer-facing tools to a Brand Agent website chatbot, including lead capture, page navigation, web search, and selected app connector actions. Brand Agent connectors should be intentionally limited. Public agents should only receive tools that are safe for visitors and useful for the customer experience. For most Brand Agents, the important tools are lead collection, page navigation, web search, and a small set of approved business actions. This is what makes Brand Agents useful as website chatbots for sales, support intake, demo routing, and customer-facing guidance. Brand Agent connector settings with built-in customer-facing tools ## Built-In Brand Agent Tools ### Lead Capture Lead Capture displays a form inside the chat when the agent needs visitor details. Use it for demo requests, sales qualification, contact collection, support intake, or any lead generation chatbot flow where your team needs clean follow-up data. Keep forms short. Ask for the information your team will actually use. Lead Capture form configuration for a Brand Agent ### Page Navigation Page Navigation turns your Brand Agent into a conversational search and navigation layer for your website. A visitor can explain what they need in natural language, and the agent can give a personalized answer while directing them to the most relevant page, such as pricing, documentation, booking, product details, or onboarding. Instead of searching through menus or guessing which page contains the answer, visitors get a website copilot that understands their intent and helps them take the next step. Use Page Navigation when the agent should help people discover and move through your site, not only answer questions inside the chat. Your website becomes easier to explore: the Brand Agent answers the visitor's specific question and routes them to the right supporting page in the same conversation. ### Web Search Web Search lets the agent look up current public information when your Knowledge Base is not enough. Use it carefully for customer-facing agents. For brand-critical answers, prefer your own Knowledge Base. ## App Connectors Brand Agents can use selected SketricGen connectors when the use case needs an external action. SketricGen supports 2,000+ connectors, but customer-facing agents should only receive the exact actions that are safe for public use. Typical customer-facing examples include: * Booking and calendar actions * CRM lead creation or updates * Support-ticket intake * Email or notification handoff * Form and spreadsheet submission * Approved ecommerce or order-status lookups Only select the exact actions the agent should use. A naming mismatch or overly broad permission can give an agent more capability than intended, so review tool access before publishing. ## Lead Review Captured leads appear in the SketricGen app. Review them alongside conversations so you can see: * What the visitor asked * Which agent response led to the lead * What details were collected * Whether the agent should ask better qualification questions Use these insights to improve the agent's instructions and Knowledge Base. For the full setup path, see [Create a Brand Agent](/brand-agents/create-brand-agent). For broader internal tools and skills, use [AI Workforce Tools, Connectors, and Skills](/user-guides/tools). # Create a Brand Agent Source: https://docs.sketricgen.ai/brand-agents/create-brand-agent Create an AI chatbot for your website with SketricGen Brand Agents: start from a URL, review Brand DNA, add knowledge, connect tools, and publish. The Brand Agent wizard helps you create a customer-facing agent from your website or from manual setup. It is the no-code path for turning your site into a branded website chatbot with knowledge, lead capture, and safe customer-facing tools. ## Start From URL or Manual Setup Choosing between website-based and manual Brand Agent setup Choose one of two creation modes: * **Website URL:** Paste your website URL. SketricGen crawls the site, extracts content, and prepares Brand DNA and Knowledge Base sources. * **Manual setup:** Enter brand details yourself and add knowledge sources manually. Use URL setup when your website already explains your product, pricing, FAQs, policies, or services. Use manual setup when the agent should be based on internal material or a new product that is not fully documented online. ## Review Brand DNA Brand Agent brand setup step Brand DNA is the agent's brand context. SketricGen can extract it from your website or let you define it manually. It can include: * Brand name * Agent name * Brand summary * Tone and style * Logo and color * Agent role * Instruction notes Review this carefully. It shapes how the Brand Agent introduces itself, answers questions, and stays aligned with your brand. ## Add Knowledge Brand Agent knowledge setup step Brand Agents need reliable information. SketricGen structures website content and other sources into an [AI Knowledge Base](/shared-platform/knowledge-bases) the agent can use during conversations. Add knowledge through: * Crawled website pages * Uploaded files * Pasted text * Q\&A pairs The Knowledge Base is what the agent uses to answer product, support, pricing, onboarding, and policy questions. If it is incomplete, the agent may need to ask clarifying questions or say that information is not available. ## Connect Optional Tools Brand Agent connector settings with built-in tools Add only the tools the public agent needs: * **Lead Capture:** collect visitor details inside chat. * **Page Navigation:** send visitors to relevant pages. * **Web Search:** look up current public information. * **Selected app connectors:** connect approved customer-facing actions. Avoid giving customer-facing agents broad internal tool access unless the use case clearly requires it. ## Test and Publish Before publishing: 1. Ask realistic customer questions. 2. Test lead capture if enabled. 3. Check whether the agent answers from the right knowledge. 4. Confirm it refuses or redirects questions outside its scope. 5. Publish and copy the widget or fullscreen link. After publishing, use Conversations and Leads to review what visitors asked and where the agent needs improvement. For deployment, see [Deploy a Brand Agent to a Custom Website](/deploy/custom-website) or the platform-specific guides for [WordPress](/deploy/wordpress) and [Shopify](/deploy/shopify). # Brand Agents: Smart Website AI Chatbots Source: https://docs.sketricgen.ai/brand-agents/overview Build a SketricGen Brand Agent as an AI chatbot for your website: train from your site, use brand knowledge, capture leads, and publish a widget. Brand Agents are customer-facing AI agents that answer questions, qualify visitors, capture leads, and guide users through your website or customer channels. Use them when you need an AI chatbot for a website, a lead generation chatbot, or an AI sales agent that is grounded in your own brand knowledge. They are designed differently from internal AI Workforce agents because public agents need tighter security, clearer tool permissions, and a simpler deployment path. Brand Agent creation paths for website-based or manual setup Built Brand Agent editor with customer-facing knowledge and connector controls ## See a Brand Agent Live The assistant in the bottom-right corner of this documentation site is a live Brand Agent built with SketricGen. It has been published as a website chat widget so documentation visitors can ask questions and receive guidance without leaving the page. SketricGen documentation with the public Lina QnA Expert Brand Agent widget open in the bottom-right corner This is a practical example of the Brand Agent experience: the agent appears directly on a public landing page or website, answers with approved brand and documentation knowledge, and helps visitors find the right information. You can open the widget on this page to try the same experience live. ## What Brand Agents Do Brand Agents can: * Start from a website URL or manual brand setup. * Extract Brand DNA from your website, including tone, positioning, audience, and product context. * Structure your website content into a Knowledge Base. * Answer visitor questions using your content and brand voice. * Capture lead details through a form inside chat. * Navigate visitors to relevant pages on your website. * Use selected customer-facing connectors when an external action is needed. * Publish as a widget or fullscreen customer-facing experience. The goal is simple: create one input-based Brand Agent that can explain your business, guide visitors, collect qualified leads, and keep working for you on your website. Start with [Create a Brand Agent](/brand-agents/create-brand-agent), then connect safe customer-facing tools in [Brand Agent Connectors and Lead Capture](/brand-agents/connectors-and-leads). ## Why Brand Agents Are Separate Customer-facing agents should not have unrestricted access to every internal app. Brand Agents use a curated tool surface so you can decide exactly what the agent can and cannot do. This makes them better suited for websites, product pages, support intake, demo qualification, onboarding, and public customer conversations. ## Brand Agent Lifecycle 1. Start from a website URL or manual setup. 2. Review the generated Brand DNA. 3. Add or edit Knowledge Base sources. 4. Connect optional tools such as Lead Capture, Page Navigation, Web Search, or selected apps. 5. Test the agent. 6. Customize and publish the widget. 7. Monitor conversations and leads. ## Conversations and Leads After publishing, review Brand Agent conversations and captured leads in SketricGen. Use these records to see what visitors asked, which answers worked, and where your Brand DNA or Knowledge Base needs improvement. ## Best Fit Use Brand Agents for: * Website visitors asking product questions * Lead qualification and demo routing * Customer support intake * Product onboarding help * Ecommerce or SaaS FAQ assistants * Brand-safe public chat experiences Use [AI Workforce](/ai-workforce/overview) instead when your team needs internal agents for long-running work, files, research, internal systems, workflow automation, or multi-agent team operations. When you are ready to publish, see [Deploy a Brand Agent to a Custom Website](/deploy/custom-website), [Deploy to WordPress](/deploy/wordpress), or [Deploy to Shopify](/deploy/shopify). # In-Product Message Tracing + New Templates + Cleaner UI (Oct 24, 2025) Source: https://docs.sketricgen.ai/changelog/2025-10-24-message-tracing-templates-ui-integrations SketricGen update: detailed in-product message tracing for debugging agents, six new templates, major UI polish, better integration handling, richer conversation inputs, and a clearer free-plan upgrade flow. This release makes it easier to **debug agents**, **launch faster with new templates**, and enjoy a cleaner, more consistent experience across SketricGen. ## TL;DR * Added **detailed in-product message tracing** to debug agent runs end-to-end * Shipped **6 new production-ready templates** (with updated thumbnails) * Big **UI/UX polish** across Chat, Flow Editor, Agent Cards, Node Inspector, Conversations * Improved **integration + connection handling** (better status + notifications + management access) * Upgraded conversation inputs: **Markdown, JSON parsing, boolean fields**, improved state handling * Free plan users now see a clearer **upgrade path** (redirect to pricing with messaging) *** ## What’s new ### 1) In-product message tracing (debug agents end-to-end) SketricGen now provides **detailed tracing inside the product** for messages and workflow testing so you can see what happened during an agent run, including: * The user message and the agent’s internal reasoning steps (high-level) * Tool calls made by the agent * Inputs and outputs for each tool call * The final response, with the full execution path visible **Why it matters:** faster debugging, easier iteration during workflow creation/testing, and fewer “mystery failures” when agents behave unexpectedly. > Screenshot/GIF placeholder: “Message trace view showing tool calls + inputs/outputs” *** ### 2) Six new templates (launch faster) We replaced the previous template set with **six new, fully tested templates**, each with a professional thumbnail. **Why it matters:** templates reduce time-to-value—start from a working pattern instead of building every workflow from scratch. *** ### 3) UI & experience improvements (cleaner, more consistent) We shipped a broad polish pass across: * Chat Window * Flow Editor * Agent Cards * Node Inspector * Conversations Page Includes improvements to layout, spacing, padding, and typography for a more consistent UI. **Why it matters:** faster building, fewer misclicks, and more confidence when configuring agents and workflows. *** ### 4) Better connection & integration handling (less friction) Integrations now feel more reliable with: * Improved **connection tracking** for agent nodes * Better **success/error notifications** during app disconnections * Easier access to **manage connected apps** and APIs **Why it matters:** fewer broken runs caused by disconnected tools, and quicker recovery when something changes in your connected stack. *** ### 5) Conversation & input handling enhancements (better structured data) Conversation fields now support: * **Markdown formatting** * **JSON parsing** * **Boolean inputs** * Improved state management **Why it matters:** agents can process and display structured inputs more reliably—especially useful for workflows that pass data between steps. *** ### 6) Subscription & access flow update (clearer upgrades) Free plan users are now redirected to the pricing page with clearer upgrade messaging. **Why it matters:** less confusion around plan limits and a more explicit path to unlock advanced capabilities. *** ## Who this release is for * Teams building and testing workflows that need **fast debugging and clear visibility** * Builders who want to ship faster using **ready-to-use templates** * Anyone relying on **tool integrations** and needing reliable connection status * Workflows that pass structured data and benefit from **JSON/boolean inputs + Markdown** *** ## How to use (quick steps) ### Debug a workflow test with message tracing 1. Run a test message in Chat or during workflow testing 2. Open the **Trace / Run Details** view for that message 3. Review the execution timeline: * tool calls * inputs/outputs * intermediate steps 4. Fix prompts/nodes/connections and re-run Screenshot showing the Trace view opened from a test run with agent steps, tool calls, and execution details ### Start from the new templates 1. Go to **Templates** 2. Pick one of the new templates (look for the updated thumbnails) 3. Duplicate/customize for your use case 4. Connect tools/apps and run a test message ### Use structured conversation fields (JSON/boolean/Markdown) 1. Open the relevant conversation input configuration 2. Choose the expected type (JSON / boolean) or add Markdown 3. Test the workflow to confirm parsing and display behavior *** ## FAQ **What is “message tracing” in SketricGen?**\ A run-level view that shows how an agent responded, including tool calls and the inputs/outputs between each step, so you can debug quickly. **Can I see tool call inputs and outputs during testing?**\ Yes—traces show the tool calls made by the agent along with what was sent and what came back. **Is tracing available during workflow creation and testing?**\ Yes—tracing is designed to help during workflow building and test runs so you can iterate faster. **Do templates work out of the box?**\ They’re tested and ready to use, but you’ll still connect your own tools (apps/APIs) and customize prompts/settings. **How do I pass structured data to an agent (JSON/boolean)?**\ Use the updated conversation fields that support JSON parsing and boolean inputs so your workflow steps receive the right shape of data. **What happens when an integration disconnects?**\ You’ll now get clearer success/error notifications and improved connection tracking so you can fix issues faster. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Traces & Optimization:** [/user-guides/traces-and-optimizations](/user-guides/traces-and-optimizations) * **Docs: Orchestration and Structured Data:** [/user-guides/orchestration-and-handoffs](/user-guides/orchestration-and-handoffs) * **Docs: Managing Tools:** [/user-guides/tools](/user-guides/tools) # Iframe + Fullscreen Chat Embeds, Forced Handoff, and a Smoother Builder (Nov 3, 2025) Source: https://docs.sketricgen.ai/changelog/2025-11-03-iframe-fullscreen-embed-forced-handoff-ga-tracking SketricGen update: embed chat via iframe + fullscreen, new Forced Handoff for tighter workflow control, FlowEditor/NodeInspector upgrades, UI consistency improvements, integration reliability fixes, and performance enhancements. This release improves three things users feel immediately: **how you embed chat**, **how you control multi-agent workflows**, and **how smooth the builder feels**—plus reliability upgrades across integrations and performance. ## TL;DR * Embed chat anywhere with **iframe support** and **fullscreen** mode * New **Forced Handoff** for more predictable control in multi-agent workflows * Improved **FlowEditor + NodeInspector** for clearer configuration and validation * Cleaner UI consistency across chat/templates/integrations (loading states, spacing, interaction feedback) * Integration reliability updates for docs links and **connector environment variable fixes for Next.js builds** * Under-the-hood refactors for **better performance, stability, and error handling** *** ## What’s new ### 1) Embed chat anywhere (iframe + fullscreen) You can now embed SketricGen chat more flexibly: * **Iframe widget support** for easy drop-in embedding * **Fullscreen mode** for a cleaner end-user experience **Why it matters:** faster deployment into websites, portals, and internal tools—without building custom chat UI from scratch. *** ### 2) More predictable multi-agent control (Forced Handoff) We introduced **Forced Handoff** to help you control agent flow in workflows: * Enforce when a workflow must transfer responsibility to the next step/agent * Reduce unexpected branching or “agent drift” in complex automations **Why it matters:** more deterministic workflows, easier testing, fewer surprises in production. Screenshot showing the Forced Handoff setting enabled in a workflow node with agent configuration *** ### 3) FlowEditor + NodeInspector upgrades (clearer configuration) We refined the workflow-building experience: * Clearer configuration and validation behavior * Improved state handling and smoother interactions while editing **Why it matters:** faster building, fewer configuration mistakes, and less friction when iterating during testing. > Screenshot/GIF placeholder: “NodeInspector validation + improved editor UX” *** ### 4) UI consistency improvements (less friction across the app) We unified styling and layout across key areas: * Chat (messages + lists) * Templates page * Integrations section * Dialogs/modals (loading states, spacing, interaction feedback) **Why it matters:** better readability and fewer “UI surprises” when you’re moving quickly between building, testing, and iterating. *** ### 5) Integrations updates (docs + reliability) * Updated and standardized integration documentation links * Fixed **connector environment variable configuration** for Next.js builds **Why it matters:** fewer setup issues and faster “first successful run” when connecting tools. *** ### 6) Reliability + performance improvements (internal upgrades that help users) We shipped internal refactors across middleware, API routes, layout metadata, and build tooling, plus cleanup of redundant configs. **Why it matters for users:** fewer edge-case errors, faster UI interactions, and more stable deployments—especially as workflows get larger. *** ## Who this release is for * Teams embedding agents into **websites, apps, and internal tools** * Builders running **multi-agent workflows** that need stronger control * Developers deploying SketricGen-connected experiences with **Next.js and connector triggers** * Anyone who wants a cleaner, faster builder experience *** ## How to use (quick steps) ### Embed chat with an iframe 1. Open your agent’s deploy/share settings 2. Choose **Iframe Widget** 3. Copy the embed snippet into your site/app 4. Enable **Fullscreen** if you want a more focused experience for users ### Use Forced Handoff in a workflow 1. Open your workflow in the **FlowEditor** 2. Configure your agent transition/handoff step 3. Enable **Forced Handoff** 4. Validate in **NodeInspector** and run a test message ### Confirm integrations stay connected 1. Open your workflow and check node connection status 2. If a tool disconnects, use the updated prompts/notifications to reconnect quickly 3. Re-run a test to confirm the workflow executes cleanly *** ## FAQ **Can I embed SketricGen chat in an iframe?** Yes—this release adds iframe widget support designed for simple embedding. **Does fullscreen mode work for embedded chat?** Yes—fullscreen is supported to improve the embedded chat experience, especially for onboarding and support flows. **What is “Forced Handoff” in a workflow?** A control mechanism that enforces when a workflow must hand off responsibility, making multi-agent behavior more predictable. **I’m using Next.js with connector triggers - what changed?** Environment variable configuration issues were fixed to improve reliability in Next.js builds. **What changed in the builder experience?** FlowEditor and NodeInspector were improved for clearer configuration, better validation behavior, and smoother editing. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Embedding & Widgets:** [/deploy/custom-website](/deploy/custom-website) * **Docs: Orchestration and Structured Data:** [/user-guides/orchestration-and-handoffs](/user-guides/orchestration-and-handoffs) * **Docs: Tools & Integrations:** [/user-guides/tools](/user-guides/tools) # Affiliate Referrals, Better Embeds, and Smoother UI (Nov 7, 2025) Source: https://docs.sketricgen.ai/changelog/2025-11-07-affiliate-referrals-widget-import-ui SketricGen mini release: expanded affiliate/referral codes, improved widget embeds (fullscreen + iframe), drag-and-drop file import, cleaner chat UI, faster performance, and reliability upgrades. This mini release focuses on practical upgrades you'll notice immediately: a more complete **affiliate/referral system**, a smoother **embed + widget experience**, and UI/performance improvements across the app. ## TL;DR * Expanded **affiliate + referral codes** across login, signup, and profiles * Improved affiliate visibility: clearer balances, codes, and eligibility/terms * Added **drag-and-drop file upload** in Import Config * Improved **fullscreen widget layout**, suggested prompts, and iframe behavior * Cleaner chat styling + better loading states across key pages * Reliability/performance improvements across components and backend lookups *** ## What’s new ### 1) Expanded affiliate & referral system (codes that actually work everywhere) We expanded referral + promo code support so it’s now available across: * Login * Signup * User profiles Affiliate users also get improved visibility into: * Balances * Codes * Terms/eligibility messaging (refined on the Affiliate page) **Why it matters:** easier attribution, fewer “did my code apply?” moments, and a cleaner experience for partners sharing SketricGen. Screenshot showing the Affiliate Dashboard with referral code, available balance, and lifetime earnings *** ### 2) Better file import & widget experience (less friction, faster testing) **Import Config** * Added **drag-and-drop** file upload for faster setup **Chat Widget / Embeds** * Multiple improvements to **fullscreen widget layout** * Better **suggested prompts** * More reliable **iframe behavior** **Why it matters:** quicker onboarding, smoother embeds on your website/app, and faster iteration when testing agent experiences. *** ### 3) UI & styling enhancements (cleaner chat, better consistency) We improved consistency across the app, including: * Chat message styling (avatars, assistant bubbles, theme-based design) * Cleaner layouts for the message list, Affiliate page, and referral content * Improved loading states and interaction polish **Why it matters:** clearer conversations, less visual noise, and a more polished experience when building and demoing workflows. *** ### 4) Performance & code quality upgrades (faster and smoother) We streamlined several hooks and components (including knowledge base and prompt experiences), refactored loading/copy logic, and removed unused behavior. **Why it matters:** fewer UI hiccups and a smoother experience as projects grow. *** ### 5) Reliability improvement (faster account lookups) We added a backend optimization (a secondary index on email) to make user lookups faster and more reliable. **Why it matters for users:** fewer edge-case delays during auth/account flows and better overall responsiveness—especially as usage scales. *** ## Who this release is for * Affiliates and partners who share SketricGen using **referral/promo codes** * Teams embedding SketricGen chat via **iframe/fullscreen widget** * Builders importing configs/files and wanting **faster setup** * Anyone who wants a cleaner UI and smoother performance *** ## How to use (quick steps) ### Use your referral or promo code 1. Go to Signup or Login 2. Enter your referral/promo code if applicable 3. Confirm it appears on your profile / affiliate area (if you’re an affiliate) ### Drag-and-drop import (Import Config) 1. Open **Import Config** 2. Drag a file into the upload area 3. Review the imported config and run a quick test ### Embed the widget (iframe/fullscreen) 1. Open your agent’s deploy/share settings 2. Choose the iframe embed option 3. Test suggested prompts + fullscreen layout on your target page *** ## FAQ **Where do I find my referral code?**\ Your referral/promo code is now surfaced across profile and affiliate-related surfaces for easier access and sharing. **Do referral codes work during signup and login?**\ Yes—support was expanded to cover login, signup, and profiles. **Can I drag-and-drop files into Import Config?**\ Yes—drag-and-drop upload is now supported. **Did the embed widget change?**\ Yes—fullscreen layout, suggested prompts, and iframe behavior were improved for smoother embeds. **Why mention backend improvements in a release note?**\ Because they reduce friction you feel as a user (faster account flows, fewer edge-case delays), even if the change is internal. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Affiliate Program:** [https://www.sketricgen.ai/affiliate-program](https://www.sketricgen.ai/affiliate-program) * **Docs: Embedding & Widgets:** [/deploy/custom-website](/deploy/custom-website) # Telegram Integration, Google Sign-In, and Mobile Upgrades (Nov 17–20, 2025) Source: https://docs.sketricgen.ai/changelog/2025-11-17-telegram-google-signin-mobile-integrations SketricGen update: Telegram integration (groups + mentions), Google Sign-In onboarding, major mobile UI improvements, stronger integration management, more reliable chat servers, and stability upgrades for long-running automations. This release bundle improves how you **deploy agents to Telegram**, how quickly users **get onboarded with Google Sign-In**, and how smooth SketricGen feels on **mobile**—plus reliability upgrades across chat servers, integrations, and long-running workflows. ## TL;DR * ✅ **Telegram integration fully released** (webhooks, bot mentions, group chat support) * 🔐 **Google Sign-In** for faster onboarding * 📱 Major **mobile responsiveness** upgrades across Dashboard, AgentSpace, FlowEditor, Conversations, Profile * 🔧 Better **integration management** (cleaner controls, safer deletion, improved status handling) * 💬 More reliable chat server behavior (fewer “Something went wrong” moments) * ⏳ Better support for **long-running automations** (improved stability for heavy execution flows) * 🧩 Instagram: backend foundation completed, **user-facing rollout coming soon** (not enabled yet) *** ## What’s new ### 1) Telegram integration (now fully available) SketricGen now supports a complete Telegram deployment experience: * Webhook registration and event handling * Bot-mention handling * Group chat support * Cleaner message routing and more real-time behavior **Why it matters:** you can deploy agents to Telegram channels/groups with fewer edge cases and smoother message handling. *** ### 2) Google Sign-In + improved onboarding guidance * Added **Google Sign-In** for faster account creation and login * Improved SSO redirect logic and onboarding messaging across mobile + desktop **Why it matters:** less friction to start building and testing workflows (especially for teams onboarding multiple users). *** ### 3) Mobile UI upgrades (big polish pass) We improved mobile layout and responsiveness across: * Dashboard * AgentSpace * FlowEditor * Conversations * Profile Highlights include: * Better breakpoints and visual consistency * Mobile sidebar state handling * Improved input behavior on mobile (less accidental zoom) * Cleaner navigation across key screens **Why it matters:** building, testing, and monitoring workflows on mobile is now far more usable. *** ### 4) Integration management improvements (cleaner controls + safer deletion) We upgraded integration management so it’s easier to maintain connected channels/tools: * Dedicated delete controls and improved authentication flows * Cleaner configuration handling * “Soft delete” behavior and status-based filtering (so inactive integrations don’t clutter your workspace) **Why it matters:** fewer “stale integrations,” fewer accidental breakages, and faster cleanup when a tool connection changes. *** ### 5) Chat server reliability improvements (fewer failures under load) We improved reliability across Telegram and third-party integration events: * Better retry handling and validation * Improved exception handling to reduce intermittent errors (including “Something went wrong”) **Why it matters:** more consistent agent replies and fewer dropped/failed runs—especially when traffic spikes or integrations send noisy events. *** ### 6) Workflow stability for heavier automation runs We strengthened reliability for longer/heavier execution flows (e.g., complex automations triggered by webhooks), including improved timeout behavior and routing stability. **Why it matters:** fewer mid-run failures for “big” workflows and more confidence running production automations. *** ### 7) Instagram integration status (coming soon) We completed the backend foundation for Instagram integration (auth + event handling + message processing).\ **Important:** the user-facing UI/workflow is **not enabled yet**. We’ll roll it out in a later release once the frontend is ready. **Why it matters:** we’re close—this groundwork is what enables a stable Instagram release rather than a fragile “beta” integration. *** ### 8) Performance & maintainability work (internal, but you’ll feel it) We shipped a set of platform refactors and cleanup work that improves long-term reliability and reduces UI friction. **Why it matters for users:** faster interactions, fewer edge-case bugs, and smoother iteration as your workspace grows. *** ## Who this release is for * Teams deploying agents on **Telegram** (support, community, sales, onboarding) * Anyone who wants **faster onboarding** (Google Sign-In) * Builders who need SketricGen to work well on **mobile** * Users managing multiple integrations and wanting **cleaner, safer integration controls** * Teams running **heavier / longer workflows** and needing more stability *** ## How to use (quick steps) ### Connect Telegram and run a test 1. Go to **Integrations** → Telegram 2. Connect your bot and complete webhook setup 3. Add the bot to a group (if needed) and mention it to trigger responses 4. Send a test message and confirm the agent replies correctly ### Sign in with Google 1. Log out (if already logged in) or open the login page 2. Choose **Continue with Google** 3. Complete onboarding steps and create your first agent/workflow ### Clean up an integration safely 1. Go to **Integrations** 2. Open the connected app/channel settings 3. Use the delete/disable controls 4. Re-run a test workflow to confirm everything is still connected and stable *** ## FAQ **Does SketricGen support Telegram group chats?**\ Yes—Telegram integration now supports group chat behavior, including bot mentions and cleaner routing. **How do I trigger the bot in a Telegram group?**\ Mention the bot (and/or follow your configured trigger behavior) to route messages to your agent. **Is Google Sign-In available for everyone?**\ Yes—Google Sign-In is now available to speed up onboarding. **Why do some changes sound “internal”?**\ Because reliability improvements (retries, validation, stability) directly reduce failed runs and make agents more consistent in production. **Is Instagram integration available now?**\ Not yet. The backend is ready, but the user-facing UI is not enabled. We’ll announce it when it’s available for users. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Telegram Integration:** [/deploy/telegram](/deploy/telegram) * **Docs: Managing Integrations:** [/user-guides/tools](/user-guides/tools) # Better Conversations, Attachments, and New Integrations (Nov 26, 2025) Source: https://docs.sketricgen.ai/changelog/2025-11-26-conversations-contacts-attachments-instagram-zapier SketricGen update: faster conversation search with Contact IDs, smoother infinite scrolling, richer messages with attachments, improved Instagram UX, Zapier + Instagram enabled in the UI, and login/UX upgrades. This release improves the day-to-day experience for teams running multi-channel agents: **finding the right conversation**, **handling richer messages**, and **managing integrations** with less friction—plus some important platform upgrades behind the scenes. ## TL;DR * Conversations now support **Contact IDs** (generation, search, filtering) for faster thread retrieval * More accurate conversation views with **platform-specific fetching** and refined filters * New controls to **disable/enable contacts** as conversation sources * Smoother infinite scroll with better pagination and a clearer “has more” state * Messages now support an **assets field** (foundation for attachments/uploads) * Instagram UX improvements + **Instagram and Zapier enabled in the UI** (with updated plan/config handling) * Login improvements: **sign out** + better redirects, plus UI cleanup and clearer messaging *** ## What’s new ### 1) Conversations are easier to find and manage (Contact IDs + better filters) Conversations now include **Contact IDs** to make identifying and retrieving threads faster: * Contact ID generation * Search and filtering by Contact ID * Platform-specific fetching for more accurate views * Improved agent-based fetching and general cleanup on the Conversations page You can also **disable/enable contacts** as conversation sources for better control over what shows up. **Why it matters:** less time hunting for the right thread, fewer “wrong platform” views, and more control over noisy conversation sources. Screenshot showing the Conversations page with platform filter dropdown displaying options including Playground, Scheduled, Public API, WhatsApp, Slack, Telegram, and Instagram *** ### 2) Smoother scrolling and pagination (less jank, more speed) We improved pagination behavior with updated limits and a clearer “has more conversations” handling to make infinite scrolling more reliable. **Why it matters:** easier triage and monitoring when you have lots of threads—especially on mobile. *** ### 3) Richer messages (attachments foundation) We updated the message schema to include an **assets field**, and standardized how Contact IDs propagate across supported messaging platforms. **Why it matters:** this is the groundwork for richer message types (attachments/uploads) across channels and more consistent cross-platform behavior. *** ### 4) Public asset uploads (in development — not fully released yet) We started building a **public asset upload system** (still in development, not fully enabled for end users yet), including: * A dedicated public assets model with authorization rules * A new upload service for more reliable file handling and error management * Platform improvements to support agent-based asset retrieval **Why it matters for users:** once completed, this enables smoother file/asset handling for agent experiences (e.g., uploading and referencing documents/media during workflows) with better reliability and security. > Note: We’ll announce this as a user-facing feature when the UI and full workflow are ready. *** ### 5) Integration updates (Instagram UX + Zapier enabled) * Improved Instagram integration UX with clearer feedback (e.g., notifications during actions) * Added platform support needed for **human takeover** logic in multi-channel agents * Enabled **Instagram** and **Zapier** in the UI with updated plan limits and configuration handling **Why it matters:** faster setup, clearer integration status, and better support for multi-channel agent operations. Screenshot showing the Integrations page with various integration options including Slack, WhatsApp, Telegram, Instagram, Zapier, Framer, WordPress, Wix, and Squarespace with Connect buttons *** ### 6) Authentication & UI improvements (small changes, big friction reduction) * Added **sign-out** support * Improved login redirects * Clarified messaging for FileSearch badge * General UI cleanup and consistency improvements **Why it matters:** fewer confusing login loops and a cleaner experience across the app. *** ## Who this release is for * Teams managing lots of customer conversations and needing **faster search + better filters** * Multi-channel agent setups where consistency across platforms matters * Anyone preparing to use richer message types (attachments/uploads) * Builders connecting Instagram or Zapier and wanting a smoother setup experience *** ## How to use (quick steps) ### Find a conversation faster with Contact IDs 1. Open **Conversations** 2. Use search/filter to query by **Contact ID** 3. Apply platform filters for more accurate views 4. Disable a contact source if it’s noisy or irrelevant ### Enable Instagram or Zapier 1. Go to **Integrations** 2. Select **Instagram** or **Zapier** 3. Follow the connect flow and confirm the connection status 4. Run a test message/workflow to verify end-to-end behavior ### Prepare for attachments (assets field) 1. If your workflows depend on file/asset content, keep an eye on upcoming “public asset uploads” 2. For now, rely on your current asset/document handling flows 3. When released, you’ll be able to attach/upload more seamlessly across channels *** ## FAQ **What is a Contact ID in SketricGen Conversations?**\ A stable identifier that helps you search, filter, and retrieve conversation threads more reliably—especially across multiple platforms. **Can I disable a contact from showing up in Conversations?**\ Yes—you can disable/enable contacts as conversation sources to reduce noise. **Are attachments supported now?**\ This release adds the schema foundation (“assets” field). User-facing attachment workflows will expand as the public asset upload system is completed and enabled. **Is the public asset upload feature available today?**\ Not fully—it’s in development. The platform foundation is in place, and we’ll announce it when the UI and full workflow are ready. **Are Instagram and Zapier available in the UI now?**\ Yes—both are enabled in the UI, with improved configuration handling and updated plan limits. **What is “human takeover” support?**\ A capability needed for multi-channel agents to hand off a conversation to a human operator when appropriate (more improvements coming). *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Conversations & Insights:** [/user-guides/conversations-and-insights](/user-guides/conversations-and-insights) * **Docs: Instagram Integration:** [/deploy/instagram](/deploy/instagram) * **Docs: Zapier Integration:** [/deploy/zapier](/deploy/zapier) # File Upload in Chat Widget, Guided Tours, and Advanced Agent Controls (Dec 13, 2025) Source: https://docs.sketricgen.ai/changelog/2025-12-13-chat-widget-file-upload-guided-tour-advanced-agent-settings SketricGen update: drag-and-drop file uploads in the chat widget (PDFs, images, text files), a guided onboarding tour, a new Space setup flow, advanced agent controls (temperature/top-p/verbosity), plus editor and MCP tool improvements. This release is about getting you from **"new workspace" → "working agent"** faster, while unlocking richer conversations (file uploads) and more precise agent behavior (advanced settings). ## TL;DR * 📎 **Upload files directly in the chat widget** (drag & drop, previews, validation) * 🧭 New **guided tour** that walks new users through building and testing a workflow * 🧱 New **Space setup modal** for faster “first agent space” creation (templates + video tutorial + empty AgentSpace) * 🎛️ **Advanced agent settings** (temperature, top-p, verbosity, reasoning effort) * 🧩 **Python SDK** released for running workflows programmatically * ✍️ Workflow editor upgrades (validation, node UI, warnings, tool node configuration) * 🔌 MCP tools improvements (better discovery, validation, connection reliability) * 📈 Better in-product stats (message counts, cost tracking, improved conversation analytics) *** ## What’s new ### 1) File upload in the chat widget (richer conversations) You can now **drag & drop files directly into the chat input** (including embedded widget use cases): **Supported file types** * Images * PDFs * Text formats: `txt`, `md`, `json`, `xml`, `csv` **What you’ll see** * Thumbnail previews before sending * Clear validation for size/type limits * Upload progress + error handling so you know what happened **Why it matters:** your agents can work with real inputs (docs, spreadsheets, screenshots, structured JSON) instead of copy/paste—making support, onboarding, and internal workflows more useful. > Screenshot/GIF placeholder: “Drag & drop upload → preview → send in widget chat” *** ### 2) Guided tour system (onboarding that actually helps) New users now get an interactive guided tour that walks through: * The palette * Agent nodes * Testing runs * Publishing/deploying Tour progress is saved so users can resume (or skip if returning). **Why it matters:** fewer “where do I start?” moments, faster activation, and less time spent guessing how to build a workflow. > Screenshot/GIF placeholder: “Guided tour steps highlighting key UI” *** ### 3) Space setup modal (faster first setup) Creating a new space is now guided with: * Template selection (carousel) * Video tutorial integration * “Empty AgentSpace” for custom builds * Auto-creation flow for brand-new users **Why it matters:** your first agent space is the hardest step—this reduces setup friction and gets users to a working workflow faster. > Screenshot/GIF placeholder: “Space setup modal with template carousel + video” *** ### 4) Advanced agent settings (more control, better outputs) We added advanced controls so you can fine-tune agent behavior per node: * Temperature (slider) * Top-P (slider) * Verbosity (low/medium/high) * Reasoning effort (minimal/low/medium/high) for supported reasoning models * Advanced mode toggle * Settings persist per agent node **Why it matters:** you can tune agents for different jobs—creative drafting vs. strict extraction vs. tool-heavy workflows—without changing the entire system. > Screenshot/GIF placeholder: “Advanced settings panel on an agent node” *** ### 5) Python SDK (run workflows programmatically) We released a **Python SDK** for running SketricGen workflows from your own systems, with smoother integrations and file-handling support. **Why it matters:** teams can trigger workflows from internal services, scripts, and products—without rebuilding logic. > Note: Public docs link should point to the SDK install + quickstart. *** ## Enhancements (you’ll feel these while building) ### Workflow editor improvements * Clearer node display (Trigger node shows “Input”) * Collapsible sections in the editor panel * Better validation errors and connection warnings * Sticky node improvements + better color handling * Tool node configuration UI improvements **Why it matters:** faster building, fewer configuration mistakes, clearer debugging during tests. *** ### Agent space upgrades * Improved workflow listing (filtering + search) * Better loading/error states * Better first-time auto-creation flow **Why it matters:** less time navigating, more time iterating. *** ### MCP tools improvements (reliability + discovery) * Default MCP type updated (more reliable streaming behavior) * Better tool discovery, validation, and error handling * Improved MCP server connection logic **Why it matters:** fewer “tool not found” moments and more consistent connectivity during runs. *** ### Stats & analytics improvements (user-facing visibility) We improved in-product stats and tracking, including: * Message count tracking for agents * Better conversation analytics * Better cost tracking at the conversation level * Improved conversation detail support **Why it matters:** clearer visibility into usage, performance, and cost drivers as you scale. *** ## Who this release is for * Teams embedding SketricGen chat and wanting **file uploads** (support, onboarding, internal tools) * New users who want a **guided first build** instead of guessing * Builders who need **more control over agent behavior** * Teams running tool-heavy workflows who want **more reliable MCP discovery and execution** * Developers who want to run workflows via **Python** *** ## How to use (quick steps) ### Upload a file in the chat widget 1. Open chat (in-app or embedded widget) 2. Drag & drop a file into the input (or select upload) 3. Preview the file and confirm it passes validation 4. Send the message and let the agent process it ### Run the guided tour 1. Create a new workspace/space (or open onboarding) 2. Start the guided tour 3. Follow the steps: build → test → publish 4. Skip anytime (returning users won’t be forced through it) ### Create a new Space using the setup modal 1. Click **Create Space** 2. Choose a template (or pick Empty Canvas) 3. Watch the quick video if needed 4. Auto-create and start testing immediately ### Tune an agent with advanced settings 1. Open the agent node in the workflow 2. Enable **Advanced mode** 3. Adjust temperature/top-p/verbosity (and reasoning effort if available) 4. Run a test message and iterate ### Use the Python SDK 1. Install the SDK (see docs) 2. Authenticate 3. Trigger a workflow run from your Python app/script 4. Use results in your system *** ## Hotfixes *No hotfixes appended to this release yet.*\ (If hotfixes happen later, append here as: **Date — change (impact)**.) *** ## FAQ **What files can I upload in the chat widget?**\ Images, PDFs, and text formats like txt/md/json/xml/csv. You’ll see validation for supported types and size limits. **Can users preview files before sending?**\ Yes—files show a preview (including thumbnails where relevant) before you send. **Is file upload available in embedded chat widgets?**\ Yes—this release focuses on the widget upload experience, including drag & drop. **What do advanced agent settings change?**\ They let you tune agent behavior per node (creativity vs. precision), and persist as part of the workflow configuration. **What is the guided tour for?**\ A step-by-step onboarding flow to help new users build, test, and publish their first workflow quickly. **Who is the Python SDK for?**\ Teams that want to run SketricGen workflows from their own backend services, scripts, or products. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Getting Started:** [/getting-started/quickstart](/getting-started/quickstart) * **Docs: Agents:** [/user-guides/agents](/user-guides/agents) * **Docs: Python SDK:** [/dev-guide/python-sdk](/dev-guide/python-sdk) # Billing Fixes, Integration Controls, and Better Onboarding (Dec 15–17, 2025) Source: https://docs.sketricgen.ai/changelog/2025-12-15-billing-fixes-integrations-onboarding SketricGen hotfix release: critical billing integrity fixes, conversation history controls for integrations, improved onboarding experience with guided tours, smoother workflow and template navigation. This hotfix release addresses critical billing issues, adds user-requested integration controls, and improves the onboarding experience—ensuring reliable payments, better conversation management, and smoother navigation. ## TL;DR * 🔒 **Critical billing fix:** prevents unintended premium access when payments fail * 💬 **Conversation history controls** for WhatsApp, Telegram, and Instagram * 🎯 **Improved onboarding** with a new guided tour system * 📝 **Smoother workflow creation** from templates (skips unnecessary setup steps) * 🔧 **Better draft detection** for workflows *** ## What's Fixed & Improved ### 1) Critical billing integrity fix (Dec 15) We fixed a critical issue where users could receive paid plan benefits despite failed Stripe payments: * Users are **no longer upgraded to paid plans** when Stripe marks a subscription as incomplete or incomplete\_expired (initial payment never succeeds) * Only **active and trialing** subscriptions now trigger plan/credit updates * Failed payments and expired subscriptions are now handled correctly, ensuring users remain on their correct plan (e.g., Free plan if payment fails) **Why it matters:** ensures billing accuracy and prevents unintended premium access, protecting both users and the platform from billing errors. *** ### 2) Conversation history controls for integrations (Dec 17) You can now **enable or disable conversation history** for messaging integrations: * Available for **WhatsApp**, **Telegram**, and **Instagram** * Controlled per integration in your integration settings * When disabled, agents won't retain conversation context between messages **Why it matters:** gives you control over data retention and agent behavior across different channels—useful for privacy requirements or when you want stateless interactions. *** ### 3) Improved onboarding experience (Dec 17) We upgraded the onboarding system with a better guided tour: * Replaced the previous onboarding with a **new guided tour system** for smoother, more reliable tours * Clearer step-by-step guidance for new users * Better visual indicators and flow **Why it matters:** faster time-to-value for new users and fewer "what do I do next?" moments. *** ### 4) Smoother workflow creation from templates (Dec 15) Creating workflows from templates is now faster: * **Templates now open directly in the workflow editor** (skips the setup modal) * Draft workflows are properly detected and labeled * Cleaner navigation flow when working with templates **Why it matters:** less clicking, faster iteration when starting from templates. *** ### 5) Knowledge base limits updated (Dec 17) For new users, the default knowledge base limit has been adjusted: * Changed from **5 knowledge bases → 1 knowledge base** for Free plan users * Aligns with updated plan structure **Why it matters:** clearer expectations for Free plan users and more consistent plan tiers. *** ## Who This Release is For * Anyone using **paid plans** (ensures billing accuracy) * Teams deploying agents on **WhatsApp, Telegram, or Instagram** who want conversation history control * **New users** going through onboarding * Builders who frequently use **templates** to create workflows * **Free plan users** (clarified knowledge base limits) *** ## How to Use (Quick Steps) ### Manage conversation history for integrations 1. Go to **Integrations** 2. Open settings for your WhatsApp, Telegram, or Instagram integration 3. Toggle **conversation history** on or off based on your needs 4. Save and test with a message to confirm behavior ### Start with the new onboarding tour 1. Create a new account or access the onboarding tour from your dashboard 2. Follow the guided steps to create your first agent/workflow 3. Complete the tour to unlock best practices and tips ### Create workflows from templates faster 1. Go to **Templates** 2. Select a template you want to use 3. Click to create from template—you'll land directly in the workflow editor 4. Customize and publish *** ## FAQ **What was the billing issue?**\ In rare cases, users could receive paid plan benefits even when their initial payment failed or expired. This has been fixed to ensure billing accuracy. **Will this affect my existing paid subscription?**\ No—this fix only prevents incorrect upgrades. If you have an active, valid subscription, nothing changes. **Can I disable conversation history for all my integrations?**\ Yes—you can control this setting individually for WhatsApp, Telegram, and Instagram integrations. **Why would I want to disable conversation history?**\ For privacy requirements, compliance, or if you want each message to be treated independently without context from previous messages. **What happened to the old onboarding tour?**\ We replaced it with a better guided tour system that's more reliable and provides a smoother experience. **Do templates still work the same way?**\ Yes—but now when you create a workflow from a template, you skip the setup modal and go directly to editing, which is faster. **What if I'm on the Free plan and had more than 1 knowledge base?**\ This change only affects new users. Existing users keep their current knowledge bases. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: WhatsApp Integration:** [/deploy/whatsapp](/deploy/whatsapp) * **Docs: Telegram Integration:** [/deploy/telegram](/deploy/telegram) * **Docs: Instagram Integration:** [/deploy/instagram](/deploy/instagram) * **Docs: Getting Started:** [/getting-started/quickstart](/getting-started/quickstart) # Faster Building: Editable Nodes, Image Generation, and Smoother Performance (Dec 29, 2025) Source: https://docs.sketricgen.ai/changelog/2025-12-29-node-inspector-editing-image-gen-integrations-performance SketricGen update: edit node instructions inline, improved advanced agent settings accuracy, image generation inside the builder, better integration handling with plan-aware alerts, and faster navigation with smoother loaders. This release makes the workflow builder feel faster and more flexible: edit node instructions inline, generate images directly inside your workflow setup, and move through the app with smoother loading and navigation. ## TL;DR * Edit **node instructions and labels inline** in Node Inspector (“click to edit”) * Improved **advanced agent settings accuracy** (model-specific configuration handling) * **Image Generation tool** is now available inside Node Inspector (when your workflow needs it) * Better integrations setup: supports API-based integrations, multiple integrations, and plan-aware upgrade alerts * Faster navigation with prefetching + skeleton loaders (less waiting, fewer spinner screens) * UI polish across onboarding, tooltips, palette, and message list * Fixes for stability and onboarding tour alignment *** ## What’s new ### 1) Edit nodes inline (less clicking, faster iteration) Node Inspector now supports: * **Editable instructions** * **Editable labels** * Clear “(click to edit)” guidance so it’s obvious where to change things **Why it matters:** faster workflow iteration—tweak copy, labels, and behavior without bouncing between screens. *** ### 2) More accurate advanced agent configuration (model-aware) We improved how advanced mode is handled so model-specific configuration is applied more accurately. **Why it matters:** fewer “why did the model behave differently?” moments and more reliable results when you’re tuning agents for different tasks. *** ### 3) Image generation inside the builder (when your workflow needs visuals) We integrated an **Image Generation tool** into the node inspector so you can add and configure it when your workflow needs visuals. **Why it matters:** if your workflow produces visual outputs (marketing images, thumbnails, product mockups, social creatives), you can configure and test image generation directly while building—without switching contexts. Screenshot showing the Edit Tool interface with Image Generation tool option highlighted, displaying various tool types including Sketric App Marketplace, File Search, Web Search, Code Interpreter, Image Generation, API Request, and Custom MCP *** ### 4) Integrations & triggers upgrades (cleaner setup + clearer plan limits) Integration setup got smarter: * **API-based integrations** supported with updated documentation links * Plan-aware messaging: **upgrade alerts** when an integration is disabled due to plan limits * Better handling for workflows with **multiple integrations** * When removing trigger integrations, you can target specific platforms so only the right connection is removed **Why it matters:** fewer broken workflows caused by misconfigured integrations, and clearer guidance when plan limits affect a node. *** ## Performance improvements (you’ll feel this everywhere) We reduced “waiting time” across common screens: * Prefetching on Agent Cards for faster navigation * Streamlined loading in AgentSpace, FlowEditor, Dashboard, Pricing, Templates * Replaced global loaders with **skeleton-based loaders** for smoother UX **Why it matters:** faster movement through the app, better perceived performance, and less context switching while building. *** ## UI/UX improvements (small friction reducers) * Loading overlays and refined opacity for clearer “fetching” feedback * Improved onboarding experience support * Better tooltips, labels, and responsiveness across nodes, palette, and message list * Password strength popover during signup **Why it matters:** fewer confusing UI moments, faster onboarding, and better clarity when configuring workflows. *** ## Fixes (stability + correctness) * Prevented runtime errors from undefined assistant messages * Fixed onboarding tour alignment so steps line up correctly * Improved user name fallback (uses email when name isn’t available) **Why it matters:** fewer random interruptions, and a more dependable onboarding experience. *** ## Internal improvements (included because they improve reliability) We removed unused components/assets, simplified loading state logic, and cleaned up styling/imports. **Why it matters for users:** fewer edge-case bugs, faster UI rendering, and a more stable platform as we ship faster. *** ## Who this release is for * Builders who want to iterate quickly on workflows with **less clicking** * Teams using **advanced agent settings** and needing more consistent behavior * Anyone adding visual generation to workflows (marketing, content, product images) * Users managing multiple integrations and wanting clearer plan-aware guidance * New users onboarding who benefit from improved clarity and responsiveness *** ## How to use (quick steps) ### Edit node instructions inline 1. Open your workflow in **FlowEditor** 2. Select a node to open **Node Inspector** 3. Click the instruction or label area (“click to edit”) 4. Save and run a quick test ### Configure image generation in a workflow 1. Open Node Inspector for the relevant step 2. Choose the **Image Generation** tool (if available in your setup) 3. Configure prompts/inputs and run a test 4. Iterate until output matches your desired style ### Fix or update integrations cleanly 1. Go to **Integrations** (or open IntegrationSection in the node) 2. If you see a plan alert, follow the upgrade guidance or switch integrations 3. If using multiple integrations, verify each is connected and configured 4. Re-run a test to confirm end-to-end behavior *** ## Hotfixes ### Hotfix — Dec 29, 2025 * **NodeInspector layout fix** — Improved styling and text wrapping so content fits within the defined width. * **Type definition standardization** — Replaced abbreviated Python type names with full names across components for clearer, consistent type handling. * **Image generation traces** — TraceDetailModal now displays generated images and relevant metadata for image-generation messages. * **StickyNode UI refinement** — Simplified layout with padding adjustments for better visual consistency. *** ## FAQ **Can I edit node instructions without opening a separate editor?**\ Yes—Node Inspector supports inline editing for instructions and labels. **Why do advanced agent settings feel different now?**\ Model-specific configuration handling was improved, so settings are applied more accurately and consistently. **Do I need image generation for every workflow?**\ No—this is for workflows that need visual outputs. If you don’t use it, nothing changes for you. **What happens if an integration is disabled due to plan limits?**\ You’ll see clearer plan-aware upgrade alerts, so you know what’s blocked and what your options are. **What’s new in traces for image generation?**\ You can now see the generated image output and metadata directly in TraceDetailModal for image-generation messages. **Why did loading screens change?**\ Skeleton loaders replace heavy global spinners to make navigation feel faster and smoother. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Tools & Integrations:** [/user-guides/tools](/user-guides/tools) * **Docs: Agents:** [/user-guides/agents](/user-guides/agents) * **Docs: Agents:** [/user-guides/agents](/user-guides/agents) # App + Landing Unified, Smoother Workflow Editing, and More Reliable Scheduling (Jan 13, 2026) Source: https://docs.sketricgen.ai/changelog/2026-01-13-app-landing-auth-flow-workflow-scheduling-integrations SketricGen update: landing page experience integrated into the app with /app routing, improved authentication flow, smoother Flow Editor panning, clearer schedule triggers, more reliable integration updates, better asset downloads, and stronger API import parsing. This release bundles two weeks of improvements focused on: a more unified SketricGen experience (site + app), smoother workflow building, and stronger reliability for scheduling, integrations, and asset handling. ## TL;DR * The **landing experience is now integrated into the app** with standardized routing and expanded public pages (blog, templates, contact, and more) * Improved **authentication flow** for cleaner redirects and more consistent login state * Flow Editor now supports **pan-on-scroll** and reduces accidental zooming * Trigger nodes now show **clear schedule badges** (including warnings when unpublished/config is missing) * More reliable integration scheduling and updates so workflows use the right connections and trigger updates are safer * Better **asset downloads** with fallback behavior when downloads are blocked * Stronger **API import parsing** (handles mixed result formats and nested JSON more reliably) *** ## What’s new ### 1) Landing + app unified (cleaner navigation and consistency) We migrated the full landing page experience directly into the application: * **Standardized routing** so app and landing live in one place * Expanded public routes: **blog**, **templates**, **contact**, and more * Unified styling and layout consistency (buttons, grids, animations, hero/CTA behaviors) **Why it matters:** everything feels like one product now—less context switching, more consistent navigation, and a smoother path from “discover” → “try” → “build.” Screenshot showing the new SketricGen landing page with dark theme, featuring "Build your team of AI Agents" headline, search bar, and suggested use cases including AI Receptionist, Personalised Outreach, Content Creation, and E-Commerce Sales Agent *** ### 2) New authentication flow (better redirects, fewer login weirdness moments) Authentication got a major reliability upgrade: * Improved login popup behavior and redirect-path handling * More consistent Google login behavior and state handling * Simplified auth logic across key surfaces (Dashboard, Templates, and landing components) Some changes are internal (new backend checks), but they reduce auth edge cases and improve consistency. **Why it matters for users:** fewer broken redirects, fewer “why am I logged out?” moments, and a smoother onboarding/login experience. *** ### 3) Smoother workflow editing (pan-on-scroll + better control) Flow Editor now supports: * **Pan-on-scroll navigation** for smoother AgentSpace movement * Reduced accidental zooming by disabling zoom-on-scroll * Improved interaction consistency for larger workflows **Why it matters:** faster workflow editing, less frustration when navigating large AgentSpaces, and better control when building quickly. *** ### 4) Trigger & scheduling clarity (badges that actually tell you what’s wrong) Trigger nodes now have clearer schedule badge logic: * **Schedule triggers** show a badge so you can see which nodes are on a schedule * Warnings appear for **unpublished schedule triggers** or when configuration is missing * Published schedules display the configured **interval or cron** values **Why it matters:** you can spot misconfigured schedules immediately and avoid silent “nothing is running” situations. *** ### 5) Integrations & webhooks reliability (fewer broken runs) Scheduling and trigger integrations are more reliable: * Scheduled webhook processing runs only on **active integrations**, so stale or inactive connections don't trigger runs * Clearer errors when no active integration is available * **Safer trigger updates** so you can change trigger integrations without creating duplicates or losing settings **Why it matters for users:** fewer workflow failures due to stale connections, and cleaner updates when you change triggers. *** ### 6) Asset handling improvements (downloads that don’t fail silently) We improved chat asset downloads: * HTTP response validation * Timestamped filenames (more unique, easier to organize) * Automatic fallback to **open in a new tab** if downloads are blocked by CORS/network behavior **Why it matters:** you can still access assets even when the browser blocks direct download. *** ### 7) API import robustness (better parsing, clearer errors) API import is now more resilient when configs are generated or returned in different formats: * Supports results as **string or object** * Better nested JSON parsing (including JSON embedded in a response field) * Clearer user-facing errors and deeper debug logging **Why it matters:** fewer broken imports and faster recovery when an imported config isn’t perfectly formatted. *** ### 8) Workflow & template enhancements (limits + cleanup messaging) We improved workflow/template behavior: * Asynchronous cleanup for deleted workflows, with clearer messaging that it’s processing in the background * Added additional template metadata support (e.g., markdown URL fields) * Plan-based workflow limit checks with better UI feedback and error handling **Why it matters:** clearer limits, clearer feedback, and fewer confusing “why is this still here?” moments after deletions. *** ## Who this release is for * New users who want a smoother path from public pages into the app experience * Teams building larger workflows who need **faster, more controlled editor navigation** * Anyone using scheduled triggers and wanting **clearer schedule status** * Builders relying on connector triggers who want **cleaner updates and fewer duplicates** * Users downloading files/assets from chat and needing reliable access * Teams importing API configs and wanting fewer parsing failures *** ## How to use (quick steps) ### Navigate the unified app experience 1. Visit public pages (e.g., /templates, /blog) 2. Click into the app experience via `/app/` 3. Use the unified header/CTA navigation to move between discovery and building ### Confirm your schedule trigger is configured 1. Open your workflow in Flow Editor 2. Select the Trigger node 3. Check the schedule badge status: * If it warns about unpublished/config-required, fix it before relying on the schedule 4. Publish and confirm interval/cron is displayed ### Update a connector trigger without duplicates 1. Open your integration/trigger settings 2. Update the trigger configuration 3. Re-run a test and confirm there are no duplicate integration records and the workflow triggers correctly ### Download chat assets reliably 1. Open a conversation with assets 2. Download the file 3. If download is blocked, use the auto “open in new tab” fallback to access it *** ## Hotfixes ### Hotfix — Jan 19, 2026 * **Improved authentication flow** — Streamlined redirect logic for more consistent navigation after login/signup. * **Better theme handling** — Updated theme switching for more reliable dark/light mode transitions. * **Category selection fix** — Blog categories now default to "All" when no specific match is found, preventing empty views. * **Workflow deletion feedback** — Improved success messages and completion handling when deleting workflows. * **Navigation improvements** — Fixed navigation behavior in Space Setup Modal for smoother onboarding. *** ## FAQ **Why did the site URLs change (what is `/app/`)?** Routing was standardized so the landing and app experience feel unified, with consistent navigation and public paths. **How do I know if my schedule trigger will actually run?** Check the schedule badge on the Trigger node—warnings appear if it’s unpublished or missing configuration, and published schedules show interval/cron values. **Will updating my connector trigger create duplicates?** This release improves trigger updates so existing integration records can be updated safely without creating duplicates. **Why are some changes described as “internal”?** Because reliability improvements (integration filtering, parsing, auth checks) reduce failures and edge cases you experience—even if the change is behind the scenes. **What if a file download is blocked in the browser?** SketricGen now falls back to opening the asset in a new tab so it stays accessible. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Tools & Integrations:** [/user-guides/tools](/user-guides/tools) * **Docs: Getting Started:** [/getting-started/quickstart](/getting-started/quickstart) # Smoother Onboarding, Infinite Scroll, and Better Tool Management (Jan 21, 2026) Source: https://docs.sketricgen.ai/changelog/2026-01-21-onboarding-infinite-scroll-tool-management SketricGen update: streamlined AgentSpace onboarding, infinite scroll for blogs and templates, improved tool enable/disable controls, better search and form components, and enhanced pricing page organization. This Mini release focuses on reducing friction for new users, improving content discovery with infinite scroll, and making tool management clearer—plus polish across key UI components. ## TL;DR * **Smoother AgentSpace onboarding** with improved new user workflow and modal interactions * **Infinite scroll** for blog posts and templates (no more pagination clicking) * **Tool enable/disable controls** for cleaner tool management * **Better search and selection** with descriptions shown so you can choose options more easily * **Improved forms and loading states** across Contact page and header * **Pricing page reorganized** for better clarity *** ## What's new ### 1) Streamlined AgentSpace onboarding (less friction for new users) We improved the new user experience in AgentSpace: * Cleaner workflow handling when starting from scratch * Better modal interactions and tour guidance * Responsive tour target styling that adapts to screen size * Dynamic layout updates for smoother onboarding flow **Why it matters:** new users get to their first workflow faster with fewer confusing steps or "what do I do next?" moments. *** ### 2) Infinite scroll for blog and templates (no more pagination) Blog posts and template listings now load automatically as you scroll: * **Infinite scrolling** with smooth pagination * Loading indicators when fetching more content * Better performance for browsing large catalogs **Why it matters:** faster content discovery—browse blogs and templates without clicking "Next Page" repeatedly. *** ### 3) Tool enable/disable management (clearer control) Tool management got simpler: * **Enable/disable toggle** for each tool so you can turn tools on or off without removing them * Better visual feedback for tool status **Why it matters:** easier to manage which tools are active in your workflows without deleting and re-adding them. *** ### 4) Improved search and selection (descriptions + better layout) Search and selection now show: * **Item descriptions** for better context when choosing options * Improved layout and styling * Clearer visual hierarchy **Why it matters:** make better choices when selecting integrations, tools, or workflow components—especially when options look similar. *** ### 5) Forms and loading states upgraded (smoother interactions) We improved form behavior and loading feedback: * Better loading state handling in Contact page and Header * Streamlined form validation * Reduced hydration warnings for faster page loads **Why it matters:** forms feel more responsive, fewer visual glitches, and clearer feedback when submitting. *** ### 6) Pricing page reorganized (cleaner presentation) The pricing page was refactored for better organization: * Cleaner component structure * Improved data fetching * Easier to scan and compare plans **Why it matters:** faster plan comparison and clearer understanding of what's included at each tier. *** ## Who this release is for * **New users** going through AgentSpace onboarding * Anyone who frequently **browses blog posts or templates** and wants faster discovery * Builders managing **multiple tools** and needing cleaner enable/disable controls * Teams using **search/selection components** and wanting better context * Anyone reviewing **pricing plans** and comparing features *** ## How to use (quick steps) ### Experience the improved onboarding 1. Create a new account or start a fresh AgentSpace 2. Follow the guided tour with improved modal interactions 3. Notice the responsive layout adjustments as you progress ### Browse with infinite scroll 1. Go to **Blog** or **Templates** page 2. Scroll down to automatically load more content 3. Keep scrolling—no need to click pagination buttons ### Manage tools with enable/disable 1. Go to **Manage Tools** 2. Toggle tools on/off without deleting them 3. Check tool status at a glance with clearer indicators ### Use improved search selection 1. Open any SearchSelect component (integrations, tools, etc.) 2. Read the descriptions to understand each option 3. Make more informed selections with better context *** ## FAQ **What changed in the onboarding flow?**\ The AgentSpace onboarding is now smoother with better workflow handling, clearer tour guidance, and responsive layout adjustments. **Does infinite scroll work on mobile?**\ Yes—infinite scroll works across all devices and automatically adapts to your screen size. **Can I still use pagination if I prefer it?**\ The blog and templates now use infinite scroll by default. This provides a faster browsing experience without manual page navigation. **What does the tool enable/disable feature do?**\ Instead of deleting tools you're not using, you can now disable them temporarily and re-enable them later without losing configuration. **Why does the SearchSelect show descriptions now?**\ To help you choose the right option—especially when multiple tools, integrations, or components have similar names. **Did the pricing plans change?**\ No—the plans are the same, but the page organization is clearer and easier to compare. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Read Our Blog:** [https://www.sketricgen.ai/blog](https://www.sketricgen.ai/blog) * **Docs: Getting Started:** [/getting-started/quickstart](/getting-started/quickstart) * **Docs: Tools Management:** [/user-guides/tools](/user-guides/tools) # Stronger Auth, Widget Upgrades, Chat Polish, and Better SEO (Jan 30, 2026) Source: https://docs.sketricgen.ai/changelog/2026-01-30-auth-widget-chat-api-landing SketricGen update: unified auth and user tracking, widget preview and dark mode, customizable CTA and QR codes, chat and layout refinements, improved API testing, and landing page metadata plus contact form improvements. This release focuses on more reliable authentication and user tracking, a richer widget and embed experience, chat and layout polish, better API testing visibility, and improved landing pages and contact forms. ## TL;DR * **Unified authentication** with better session handling and clearer error reporting * **Usage and activity visibility**: total credits consumed, last activity date, and clearer usage tracking * **Widget & embed upgrades**: preview mode toggle, dark mode support, customizable CTA button, and QR code download for workflows * **Chat and layout refinements** for consistency, responsiveness, and cleaner input behavior * **API dry-run improvements** with better error handling and response visibility in test results * **Landing pages & contact**: metadata for SEO and social sharing, plus a simpler contact form with category selection *** ## What's new ### 1) Stronger authentication and user tracking We improved sign-in and how your usage is tracked: * More consistent session handling and clearer error reporting when something goes wrong * Your account now tracks **total credits consumed** and **last activity date** for better visibility * **Last activity date** is set automatically when you sign up **Why it matters:** fewer auth hiccups, clearer errors when something goes wrong, and better visibility into your usage and activity. *** ### 2) Widget and embed experience (preview, dark mode, CTA, QR) The widget and fullscreen embed got several upgrades: * **Preview mode toggle** to switch between widget and fullscreen views * **Dark mode support** for fullscreen and iframe widgets with dynamic theme handling * **Customizable CTA button** in fullscreen widget (title + link) * **QR code download** for workflows using workflow URLs * Cleaner widget configuration and visibility control **Why it matters:** you can preview how your widget looks in both modes, match your site’s theme with dark mode, add a custom call-to-action, and share workflows easily via QR code. *** ### 3) Chat and layout polish (consistency and responsiveness) Chat and related UI received refinements: * Improved chat layout, input area, preview, and sidebar styling for consistency and responsiveness * Better input behavior and refined avatar and input area layout * Fixed minor styling issues in fullscreen preview components **Why it matters:** chat feels more consistent, inputs behave more predictably, and the fullscreen preview looks cleaner. *** ### 4) API testing and developer experience API dry-run and test feedback are easier to use: * **Dry-run testing** with better error handling and clearer feedback * **Response bodies** are included in test results for clearer visibility * Refined test-import handling and feedback in the API test dialog **Why it matters:** when testing APIs, you see full responses and get clearer errors, so debugging and validation are faster. *** ### 5) Landing pages, SEO, and contact form Public pages and contact are easier to find and use: * **Metadata** added to authentication and landing pages for improved SEO and social sharing (preview cards, titles, descriptions) * **Contact form UX**: single name field and **request category selection** for clearer submissions * Fixed email placeholder text for better clarity **Why it matters:** your shared links look better on social and in search, and contacting support is simpler with a single name field and category choice. *** ## Who this release is for * **Users and teams** who want more reliable sign-in and clearer usage/activity tracking * **Embedders** who need preview mode, dark mode, custom CTAs, or QR codes for workflows * **Chat users** who care about consistent layout and input behavior * **Developers** testing APIs and wanting better dry-run results and error visibility * **Anyone sharing landing or auth pages** and wanting better SEO and social previews * **Support and sales** who use the contact form and want clearer categories and UX *** ## How to use (quick steps) ### Preview and customize the widget 1. Open your widget or fullscreen embed settings 2. Use the **preview mode toggle** to switch between widget and fullscreen views 3. Enable **dark mode** if your site uses a dark theme 4. Set a **custom CTA** (title + link) in the fullscreen widget 5. Use **QR code download** for a workflow to share its URL easily ### Test APIs with better visibility 1. Open **ApiDryRunDialog** or your API test flow 2. Run a dry-run; requests go through Next.js API endpoints 3. Check test results for **response bodies** and clearer error messages ### Submit a clearer contact request 1. Go to the **Contact** page 2. Enter your name in the single name field 3. Choose a **request category** and fill in the rest 4. Submit; the team gets a better-structured request *** ## FAQ **What is the “unified auth utility”?**\ We replaced Amplify-based session fetching with a single auth utility so session and attribute handling are consistent and errors are reported more clearly. **Does dark mode work in iframes?**\ Yes—dark mode is supported for both fullscreen and iframe widgets, with dynamic theme handling. **Can I still see API errors if the response is large?** Yes—response bodies are now included in dry-run test results so you can see what the API returned, which helps with debugging. The Auto Fill API Feature converts and API docs into usable tool by the Agent, just taking the documentation link as input. **What request categories are available on the contact form?** The contact form now includes a category selector so you can label your request (e.g. support, sales, feedback). Check the live form for the exact options. **Did credit deduction behavior change?** Credit logic now tracks total credits consumed with improved logging; behavior is more transparent, not stricter. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Getting Started:** [/getting-started/quickstart](/getting-started/quickstart) * **Docs: Widget & Deploy:** [/deploy/custom-website](/deploy/custom-website) * **Docs: Public API:** [/dev-guide/public-api-usage](/dev-guide/public-api-usage) # Brand Agent, Crawl Pipeline, Dashboard Feedback & Voice Input (Feb 12, 2026) Source: https://docs.sketricgen.ai/changelog/2026-02-12-brand-agent-crawl-dashboard-voice-ui SketricGen update: Brand Agent creation and workflow generation, crawl pipeline with live progress, dashboard upgrade alerts and feedback, voice input with ScribeMic, and UI/styling improvements. This release adds Brand Agent creation with workflow generation, a stronger crawl pipeline with live progress and plan-limit disclaimers, dashboard upgrade alerts and in-app feedback, voice input via ScribeMic, and UI and styling polish across the app. ## TL;DR * **Brand Agent creation** from the dashboard with progress tracking and workflow generation—see workflows build in real time and get upgrade prompts when you hit capacity * **Crawl pipeline** with live status and progress, plan-limit disclaimers up front, and secure access to crawled content * **Dashboard** upgrade alerts when you approach workflow limits, clearer permissions and validation, and **in-app feedback** from dashboard and chat * **Voice input** in chat so you can speak instead of type when it's faster * **UI and styling** improvements: refreshed buttons and badges, better dark mode, and consistent look across landing, dashboard, and chat *** ## What's new ### 1) Brand Agent creation and workflow generation We added an end-to-end flow for creating Brand Agents and generating workflows from the dashboard: * **Brand Agent creation** from the dashboard with clear progress as workflows are built * Better brand data capture and cost visibility for Brand Agent setup * Improved error handling and event handling so the dashboard reflects workflow-building progress * Simplified workflow limit checks and upgrade alerts when you hit capacity **Why it matters:** you can create Brand Agents from the dashboard, see progress as workflows are built, and get clearer feedback when you hit workflow capacity—with upgrade prompts when needed. *** ### 2) Crawl pipeline and infrastructure The crawl system is more robust and transparent: * **Live progress** and status so you see how each crawl job is doing * **Plan limit disclaimers** shown up front so you know when a crawl is affected by your plan * More reliable job processing and better error handling when creating crawl jobs * Crawled content is available via secure links; job expiration was tightened for better resource use **Why it matters:** crawl jobs run more reliably, you see live progress and plan limits up front, and you can access crawled content securely. *** ### 3) Dashboard, permissions, and feedback Dashboard and permissions are clearer; feedback is built in: * **Upgrade alerts** on the Dashboard when you approach or hit workflow capacity limits * **User permission handling** and workflow validation improved for more predictable behavior * **In-app feedback** from the dashboard and chat so you can send comments or issues without leaving the app * Clearer indicators when knowledge base or tools are processing **Why it matters:** you know when to upgrade, permissions and validation behave more predictably, and you can give feedback directly from the dashboard and chat. *** ### 4) Voice input in chat Voice is now part of the input experience: * **Voice input** in the chat flow so you can speak instead of type * Integrated voice pipeline for reliable transcription **Why it matters:** you can speak your input instead of typing when it’s faster or more convenient. *** ### 5) UI, UX, and styling Visual consistency and polish across the app: * **Button and badge** styles updated with a refreshed look and better dark mode * Landing page, hero, and chat header styling refined for consistency * Dashboard and chat component styles updated for a more consistent look **Why it matters:** the app feels more cohesive, and dark mode and key surfaces (landing, dashboard, chat) are easier on the eye and easier to use. *** ## Who this release is for * **Teams building Brand Agents** who want dashboard-driven creation and workflow generation with clear progress * **Users who run crawls** and want live status, plan-limit clarity, and reliable S3 access * **Dashboard users** who care about upgrade alerts, permissions, and in-app feedback * **Chat users** who prefer **voice input** when speaking is faster than typing * **Everyone** who benefits from clearer UI, dark mode, and consistent styling *** ## How to use (quick steps) ### Create a Brand Agent and track workflow progress 1. Use the **Brand Agent creation** flow from the dashboard 2. Watch **progress tracking** as the workflow is built 3. If you hit workflow limits, use the **upgrade alert** to see your options ### Run and monitor crawl jobs 1. Create a crawl job from the dashboard or API 2. Watch **live progress** and any **plan limit** disclaimers 3. Access crawled content via the secure links provided when the job completes ### Give feedback from the app 1. In the **Dashboard** or **chat** window, use the in-app feedback option 2. Submit feedback in context; it's sent to the team for follow-up ### Use voice input 1. In chat, use the **mic button** (where available) to speak your input 2. Voice is transcribed and sent as your message *** ## FAQ **What is Brand Agent creation?**\ It's the flow in the dashboard that lets you create a Brand Agent with your brand details and optional requirements, then generate or update workflows—with progress visible along the way. **Do crawl jobs expire sooner now?**\ Yes—crawl job expiration was reduced for better resource use; you still get secure links to access crawled content before expiry. **Where do I see workflow capacity limits?**\ The Dashboard shows **upgrade alerts** when you approach or hit workflow capacity, with clearer logic around limits and upgrade prompts. **Where can I send feedback?**\ You can send in-app feedback from the Dashboard and from the chat window—comments or issues go to the team without leaving the app. **Where is voice input available?**\ Voice input is available in the chat flow where the mic option is shown; use it to speak instead of type when it's more convenient. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Getting Started:** [/getting-started/quickstart](/getting-started/quickstart) * **Docs: Widget & Deploy:** [/deploy/custom-website](/deploy/custom-website) * **Docs: Public API:** [/dev-guide/public-api-usage](/dev-guide/public-api-usage) # Auth, Voice Polish, Limits & Dashboard UX (Feb 17, 2026) Source: https://docs.sketricgen.ai/changelog/2026-02-17-auth-utm-voice-limits-dashboard-ui SketricGen update: auth and onboarding redirect improvements, plan limit checks with upgrade CTAs, centralized voice input with auto-stop and silence timeout, dashboard and chat polish, and UI updates including blog theme and resources dropdown. This release makes auth and onboarding redirects more reliable, tightens plan-limit checks with clearer upgrade prompts, centralizes voice input with better recording behavior, and polishes dashboard, chat, and UI—including blog theme toggle and a resources dropdown in the header. ## TL;DR * **Sign-in and redirects:** after login or SSO you're taken to the right place more reliably, with fewer redirect quirks * **Plan limits:** faster limit checks and a clear **upgrade CTA** when you hit workflow or knowledge base limits; agents created in the background start inactive so they don't count until you turn them on * **Voice input:** consistent behavior everywhere—**auto-stop** and **silence timeout** so recordings end predictably; the UI clearly shows when the mic is on * **Dashboard and chat:** better mobile routing, fewer duplicate tool messages, and clearer handling when you hit plan limits * **UI updates:** blog **theme toggle**, header **resources** dropdown, improved demo CTA, and styling refinements on landing, blog, and contact pages *** ## What's new ### 1) Authentication and onboarding flow Login and post-login navigation are more predictable: * After sign-in or SSO, you're sent to the right place based on whether you've already created agents and your workflow list * Redirect logic is simpler and more consistent across login popup and SSO * Onboarding routing and post-login navigation are more stable **Why it matters:** you’re sent to the right place after sign-in or SSO more reliably, with fewer redirect quirks. *** ### 2) Plan limits and upgrade prompts Limit checks are faster and clearer when you hit caps: * **Workflow and knowledge base** limit checks run more efficiently with fewer redundant requests * **Upgrade CTA** appears when you hit workflow or knowledge base plan limits so you can see upgrade options quickly * Agents created in the background (e.g. during setup) start **inactive** so they don't count against your active limit until you turn them on **Why it matters:** you see plan limits and upgrade options without extra delay, and background-created agents don't count as active until you activate them. *** ### 3) Voice input (consistent behavior everywhere) Voice recording is unified and more reliable across the app: * **Auto-stop** and **silence timeout** so recordings end predictably when you stop speaking * Same behavior in dashboard chat and workflow chat—no surprises when you switch contexts * The UI clearly shows when the mic is on and when input is disabled (e.g. while the system is processing) **Why it matters:** voice input behaves the same everywhere, stops when you go silent or when auto-stop triggers, and the UI clearly shows when the mic is on. *** ### 4) Dashboard and chat improvements Dashboard chat and error handling are smoother: * **Mobile routing** in dashboard chat improved so navigation feels right on smaller screens * You can't send new messages while the system is processing—the input state reflects this clearly * Fewer duplicate tool messages in the thread for a cleaner conversation view * When you hit a plan limit in chat, the error is explained and an upgrade path is shown **Why it matters:** chat works better on mobile, duplicate tool messages are reduced, and hitting a plan limit in chat is explained and surfaced clearly. *** ### 5) UI and experience updates Surfaces and navigation got small upgrades: * **Blog** client **theme toggle** for light/dark * **Header** updated with a **resources** dropdown * **Demo CTA** visibility improved * Styling refinements on **landing**, **blog**, and **contact** pages **Why it matters:** you can switch blog theme, find resources from the header, and see a clearer demo CTA; landing and contact feel more consistent. *** ## Who this release is for * **New users** going through auth and onboarding who benefit from stable redirects * **Teams near plan limits** who want clear upgrade CTAs and fewer redundant limit checks * **Voice users** who want consistent mic behavior, auto-stop, and silence timeout * **Dashboard and chat users** on mobile or hitting plan limits who want better routing and error handling * **Readers and visitors** who use the blog theme toggle, resources dropdown, or demo CTA *** ## How to use (quick steps) ### Use voice input with the new behavior 1. Use the **mic button** in dashboard or workflow chat to start recording 2. Speak; recording **auto-stops** on silence timeout or when you stop 3. If the mic or input is disabled, wait for the current action to finish—the UI reflects recording state ### When you hit a plan limit 1. Look for the **upgrade CTA** when you hit workflow or knowledge base limits 2. Use it to see upgrade options; limit checks run in the background with fewer extra calls ### Use the new UI options 1. On the **blog**, use the **theme toggle** for light or dark 2. In the **header**, open the **resources** dropdown to find links 3. Use the improved **demo CTA** where it’s shown on landing or marketing pages *** ## FAQ **Why do some agents start inactive?**\ Agents created in the background (e.g. during setup or automation) now start inactive so they don't count against your active workflow limit until you turn them on. **Does voice auto-stop work everywhere?**\ Yes. In dashboard chat and workflow chat, voice recording uses auto-stop and silence timeout so behavior is consistent. The mic stops when you go silent or when auto-stop triggers. **Where is the resources dropdown?**\ In the main site **header**; use it to access resources and key links without leaving the page. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Getting Started:** [/getting-started/quickstart](/getting-started/quickstart) * **Docs: Widget & Deploy:** [/deploy/custom-website](/deploy/custom-website) * **Docs: Public API:** [/dev-guide/public-api-usage](/dev-guide/public-api-usage) # Max Agent Builder, Trial Plan & Subscription Updates (13 March 2026) Source: https://docs.sketricgen.ai/changelog/2026-03-13-max-trial-subscription-updates SketricGen update: significant updates to the Max agent builder; new Trial plan so you can test all premium features free; subscription management, conversation limits, templates, consent, customer portal, and resizable Max widget. We've introduced **significant updates to the Max agent builder** and a new **Trial plan**: you can now **officially test all premium features at no cost** with the trial. This release also brings full subscription management (update, cancel, resume), per-conversation message limits, template URL and metadata improvements, cookie consent and analytics, customer portal flows, trial status on profile, end-trial-early option, and a resizable Max widget in the flow editor. ## TL;DR * **Max agent builder & Trial plan:** **Trial plan** is here—**test all premium features free** with no payment required. Max (agent builder) gets ongoing improvements so you can build and try workflows with full access during trial. * **Subscription management:** **Update**, **cancel**, and **resume** your subscription from the **customer portal**; cancel can be immediate or at period end; your account stays in sync with billing * **Trial status & end trial early:** **Trial badges** and **trial end date** on your **profile**; option to **end your trial early** when you're ready to choose a plan or cancel * **Conversation limits:** Conversation views respect **plan-based message limits** per conversation while loading recent history smoothly * **Template URL and metadata:** Template pages use a single **template** URL pattern and **richer metadata** for SEO and sharing * **Cookie consent and analytics:** **Cookie consent** where required, with **region-based** behavior; analytics and events align with consent and user identity * **Resizable Max widget:** The **Max** (agent builder) **chat panel** in the **flow editor** is **resizable**—drag the edge to make it wider or narrower * **Trace errors & conversation list:** **Trace detail** views show errors in a clearer format; **conversation list** and message styling are refined *** ## What's new ### 1) Max agent builder & Trial plan * **Trial plan:** You can now **officially test all premium features without paying**. Sign up and use the trial to explore the full platform—workflows, integrations, and Max—at no cost. * **Max agent builder** continues to receive significant updates so you can design, test, and iterate on agent workflows with full access during your trial. **Why it matters:** you can evaluate SketricGen and Max with no commitment, then choose a plan when you're ready. *** ### 2) Subscription update, cancel, and resume You can manage your subscription from the app with clearer feedback: * **Customer portal** supports **subscription update** (change plan or price), **cancellation** (immediate or at period end), and **resume** (reactivate a cancelled subscription before it lapses) * Access the portal from **profile** and **pricing** pages; **error handling** is improved so you see clear messages if something fails * **User record** stays in sync with your billing provider so your plan and trial state are accurate everywhere in the app **Why it matters:** you can upgrade, change plan, cancel, or resume without leaving SketricGen, with clearer feedback when something goes wrong. *** ### 3) Trial status on profile & end trial early Trial handling is more transparent and flexible: * **Trial status** and **trial end date** are shown on your **profile** page with **badges** so you know when your trial ends * **End trial early** — you can choose to end your trial before it expires (e.g. to switch to a paid plan or leave) * **Plan info** in the workflow editor shows **trial messaging** and **app allowance** in a simpler, more consistent way **Why it matters:** you know exactly when your trial ends and can end it early if you want; fewer surprises about trial limits when building workflows. *** ### 4) Conversation message limits and history Conversation views now respect your plan while keeping recent history accessible: * **Per-conversation message limits** apply based on your plan so you see the right amount of history per conversation * **Recent history** still loads smoothly within those limits—no abrupt cutoffs * Conversation tracking is improved with **platform identifiers** and **unread status** for better multi-channel management **Why it matters:** you stay within plan limits per conversation while still viewing recent messages, and conversations are easier to manage across channels. *** ### 5) Template URL and page metadata Template pages are easier to find and share: * **Template URL** is standardized to a single pattern (e.g. `/template/...`) so links and bookmarks stay consistent * **Template listing and detail pages** use this structure for clearer navigation * **Template pages** use **parsed frontmatter** to drive titles, descriptions, and display details for better SEO and sharing **Why it matters:** template links are reliable, and template pages show up better in search and when shared. *** ### 6) Cookie consent and analytics Cookie consent and analytics work together while respecting privacy: * **Cookie consent** is available where required (e.g. cookie banner) so you can accept or reject categories * **Region-based consent** logic applies so behavior matches local expectations * **Consent preference** updates use a single, consistent event so consent state is applied once and correctly; fewer duplicate or noisy updates * **Analytics and events** align with consent state and user identifiers across the app **Why it matters:** the site can respect your consent choices and still provide useful analytics where you've agreed. *** ### 7) Resizable Max widget in the flow editor The Max (agent builder) chat panel in the flow editor adapts to your workspace: * **Max widget panel** in the **flow editor** is **resizable** — drag the panel edge to make it wider or narrower * Width is constrained to a comfortable range so the layout stays usable **Why it matters:** you can size the Max chat panel to fit your workflow instead of a fixed width. *** ### 8) Trace error display & conversation list * **Trace detail** dialog now shows **errors** in a **normalized, consistent format** for faster debugging * **Conversation list** item padding and **selected state** are adjusted for clearer hierarchy; **message bubbles** and list layout are refined for consistency **Why it matters:** when a run fails, you see what went wrong clearly; the conversation list and messages are easier to read and navigate. *** ## Who this release is for * **New users** who want to try SketricGen and **Max** with the **free Trial plan** and full premium features * **Subscribers** who want to update, cancel, or resume their plan from the app * **Trial users** who want to see trial end date and optionally end trial early * **Users on limited plans** who want clear, per-conversation message limits without losing recent history * **Anyone sharing or bookmarking template pages** who benefits from consistent URLs and better metadata * **Users in regions with consent requirements** who want cookie consent and analytics that respect their choices * **Flow editor users** who want a resizable Max chat panel and clearer plan/trial messaging * **Debuggers** who use trace details and want clearer error display * **Conversation users** who want a cleaner list and message layout *** ## How to use (quick steps) ### Start with the Trial plan 1. Sign up at [SketricGen](https://www.sketricgen.ai/) and start your **trial** to **test all premium features** at no cost 2. Use **Max** (agent builder) in the flow editor to design and run workflows with full access during the trial ### Manage your subscription 1. Go to **Profile** or the **pricing** page and open the **customer portal** (e.g. manage subscription link) 2. To **update:** choose a new plan or price and confirm 3. To **cancel:** start cancellation; choose immediate or **cancel at period end** 4. To **resume:** if you cancelled at period end, use **resume** before the period ends to keep your subscription active ### Check or end your trial 1. Open your **Profile** to see **trial badges** and **trial end date** 2. To **end trial early,** use the option from profile or pricing (e.g. end trial early or choose a plan) ### Resize the Max widget 1. Open the **flow editor** and show the **Max** (agent builder) panel 2. **Drag the panel edge** (left or right) to resize; the panel width stays within a comfortable range ### Work with conversation limits 1. Open **Conversations** and select a conversation; recent messages load within your **plan limits** 2. If you hit the limit, older messages may not load; upgrade your plan for higher limits ### Use template pages 1. Browse templates at the **template** URL (e.g. from the app or docs); links use the single `/template/...` pattern 2. Share or bookmark template pages; titles and descriptions come from the page metadata for better previews *** ## FAQ **Can I try premium features without paying?**\ Yes. The **Trial plan** lets you **officially test all premium features** at no cost. Use the trial to explore Max, workflows, and integrations; when you're ready, choose a paid plan or end the trial. **Where do I see my trial status and trial end date?**\ On your **Profile** page. Trial status is shown with **badges** and your **trial end date** is displayed so you know when the trial expires. **Where do I update or cancel my subscription?**\ Use the **customer portal** from your **Profile** or the **pricing** page. The portal supports subscription update, cancellation (including cancel-at-period-end), and **resume**. **Can I resume a cancelled subscription?**\ Yes. If you chose **cancel at period end**, you can **resume** from the customer portal (via profile or pricing) before the period ends. After the period ends, you'd need to subscribe again. **How do I end my trial early?**\ Use the **end trial early** option from your **Profile** or the **pricing** page. You can then choose a paid plan or leave; your trial will end as requested. **Do conversation message limits apply per conversation or in total?**\ Limits are **per conversation**. Each conversation respects the plan-based message limit; recent history within that limit loads normally. **What is the template page URL?**\ Template pages use a single URL pattern: `https://www.sketricgen.ai/template/{slug}`. Use this when linking or bookmarking. **Can I resize the Max chat panel in the flow editor?**\ Yes. The **Max** (agent builder) **panel** in the **flow editor** is **resizable**. Drag the panel edge to make it wider or narrower; the width is constrained so the layout stays usable. **Does SketricGen use cookie consent?**\ Yes, where required. The site can show a cookie consent banner and respect your choices; analytics and events align with your consent. Behavior may vary by region to match local expectations. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Docs: Getting Started:** [/getting-started/quickstart](/getting-started/quickstart) * **Docs: Widget & Deploy:** [/deploy/custom-website](/deploy/custom-website) * **Docs: Public API:** [/dev-guide/public-api-usage](/dev-guide/public-api-usage) # Practical Agents That Do Real Work Source: https://docs.sketricgen.ai/changelog/2026-05-25-practical-agents-that-do-real-work SketricGen V3 update: website Brand Agents for AI chatbots and lead capture, plus AI Workforce for automation, files, connectors, skills, traces, and models. SketricGen V3 is about moving from simple chatbots and demos to agents that can help run real business work. The product now separates two jobs that need different defaults: * **Brand Agents** for customer-facing website chatbots where security, brand behavior, lead capture, and controlled tools matter. * **AI Workforce** for internal AI automation where teams need longer runs, files, connectors, skills, traces, and broader model choice. ## Brand Agents: Customer-Facing Agents Brand Agents are built for visitors, leads, prospects, customers, and public channels. You can start from a website URL or manual setup. SketricGen extracts Brand DNA, structures website content into a Knowledge Base, and helps you publish an agent that can answer questions, guide visitors, and collect leads. Brand Agents now support: * Website-based Brand DNA extraction * Knowledge Base setup from scraped website content * Lead capture inside chat * Page navigation for guiding visitors to the right page * Web search when current public information is useful * Selected customer-facing app connectors * Conversation and lead review after publishing The main idea is simple: one business input can become a practical Brand Agent that explains your offer, answers questions, collects qualified leads, and keeps improving as you refine its knowledge. ## AI Workforce: Internal Agents for Longer Work AI Workforce is for work that happens inside your team. Agent Spaces now use a clearer structure: one main agent talks to the user and focused sub agents help with bounded tasks. Sub agents do not create their own sub agents, so runs stay easier to trace and debug. AI Workforce can now support: * Longer runs that may take minutes instead of seconds * Planning and todos during a run * Files, run workspaces, and generated artifacts * App connectors for internal systems * Reusable skills from the marketplace * AI Workforce templates * Rich traces for reviewing what happened This makes AI Workforce closer to a work harness for business tasks. It is powerful like modern coding agents, but connected to your apps, files, and internal processes instead of only a local codebase. ## Connectors and Skills SketricGen V3 expands what agents can do after they understand a task. AI Workforce can use broad connector access through 2,000+ SketricGen connectors, API Request, and custom tools. Brand Agents use a smaller curated connector surface. That keeps public agents safer and easier to approve before publishing. The skill marketplace is now a major part of building capable agents. You can search for skills, review their source, verification, categories, and usage signals, then install them directly onto an agent. Once installed, the skill gives the agent a curated way to perform that task during a run. That means you can build from proven capabilities instead of starting every agent from a blank prompt. Connect the apps, install the skill, add instructions, choose the model, and test the run. ## Knowledge Bases With More Control Knowledge Bases remain the trusted source layer for both product paths. For website scraping, SketricGen can show the crawled pages and the amount of content detected. The latest update also gives more control over scraped website content: crawled pages can be stored as Markdown so you can view, edit, and improve what the agent will later use. That matters because cleaner scraped Markdown usually produces better answers. If a website crawl includes irrelevant navigation, repeated footer text, outdated copy, or missing context, you can tighten the source before relying on it. ## Files, Artifacts, and Traces AI Workforce runs can now produce and inspect files during longer tasks. Artifacts and traces make it easier to review what happened before using the output. Use traces to check: * Which agent or sub agent handled each step * Which tools and connectors were called * What failed or slowed the run down * What files or artifacts were created * How much credit the run consumed ## Model Choice SketricGen now uses a broader model catalog instead of assuming every agent must use one provider. Available models and settings may change over time, so the in-product model picker is the current source of truth. ## Existing Workflows Existing workflows should continue to work, but the UI is now organized around the way teams actually use agents: * Brand Agents for public customer-facing deployment * AI Workforce for internal work, sub agents, tools, skills, templates, files, runs, and traces * Knowledge Bases for trusted content * Conversations and Leads for reviewing customer-facing results ## Start Here * [Brand Agents vs AI Workforce](/getting-started/choose-product) * [Create a Brand Agent](/brand-agents/create-brand-agent) * [AI Workforce Overview](/ai-workforce/overview) * [Knowledge Bases](/shared-platform/knowledge-bases) # Persistent Website Chats, Safer Uploads & Clearer Plan Limits (June 5, 2026) Source: https://docs.sketricgen.ai/changelog/2026-06-05-widget-navigation-crawl-limits-cookie-controls SketricGen update: persistent Brand Agent chats, smoother website navigation, clearer knowledge base limits, safer uploads, and improved cookie controls. This release makes customer-facing Brand Agents more dependable across page changes and reloads. It also gives you clearer knowledge base limits, safer file uploads, and more control over cookie preferences. ## TL;DR * **Persistent website conversations:** Visitors can continue a Brand Agent conversation after reloading the page. * **Smoother page navigation:** Brand Agents can guide visitors to another page without unnecessarily closing the chat. * **Safer uploads:** Brand Agent file uploads are limited to images and PDFs. * **Clearer crawl limits:** Free-plan users receive a clear upgrade prompt when a website crawl reaches the 24-page limit. * **Better cookie controls:** Visitors can accept, reject, or customize non-essential cookies and reopen their preferences later. * **More reliable verification:** Sign-up guidance now reminds users to check spam when a verification email is missing. *** ## What's new ### 1) Website conversations that survive reloads and navigation Brand Agent widgets now preserve more of the visitor's experience while they move around your website: * **Conversation restore:** A visitor can reload the page and continue the same widget conversation. * **In-site navigation:** When a Brand Agent directs someone to another page on the same website, supported sites can keep the widget open during the transition. * **Completed navigation replies:** The agent can finish and save its response before the page changes. * **Cleaner navigation prompts:** Page-navigation actions appear separately from technical tool activity, making the chat easier to follow. **Why it matters:** Visitors can explore your website without repeatedly losing context or restarting their conversation. *** ### 2) Safer Brand Agent file uploads Customer-facing upload options are now more focused: * **Supported uploads:** Brand Agent visitors can attach images and PDF documents. * **Clearer boundaries:** Other file types are excluded from the public-facing upload flow. * **More reliable loading:** Widget startup and conversation restoration handle overlapping requests more consistently. **Why it matters:** Your public agent receives useful visual and document context through a simpler, safer upload experience. *** ### 3) Clearer knowledge base crawl limits Website crawling now gives Free-plan users better feedback: * **24-page Free-plan limit:** Website knowledge base crawls on the Free plan support up to 24 pages. * **Visible limit state:** If a crawl reaches the plan limit, the interface explains what happened. * **Upgrade path:** The limit message provides a direct route to plans with higher allowances. **Why it matters:** You can understand why a crawl stopped and decide whether to narrow the source or upgrade for a larger website. *** ### 4) More useful cookie preferences Cookie consent is easier for visitors to control: * **Three clear choices:** Visitors can accept all cookies, reject non-essential cookies, or customize preferences by category. * **Saved preferences:** Choices persist for future visits. * **Cookie settings link:** Visitors can reopen their preferences from the website footer. * **Production-only analytics:** Production analytics are kept separate from staging activity. **Why it matters:** Visitors get understandable privacy controls, while site analytics respect the preferences they select. *** ## Who this release is for * **Brand Agent owners** who want website conversations to remain useful across page changes. * **Website visitors** who expect chats to survive an accidental refresh. * **Free-plan users** creating knowledge bases from websites. * **Teams accepting visitor files** through a customer-facing agent. * **Privacy-conscious visitors** who want granular cookie controls. ## How to use (quick steps) ### Test conversation restoration 1. Open your published **Brand Agent** widget and send a message. 2. Reload the page. 3. Reopen the widget and confirm that the conversation continues. ### Review a website crawl limit 1. Create or edit a **Knowledge Base** using a website source. 2. Start the crawl and monitor its progress. 3. If the Free-plan page limit is reached, follow the on-screen guidance to refine the source or review upgrade options. ### Change cookie preferences 1. Choose **Accept all**, **Reject non-essential**, or **Customize** when the cookie notice appears. 2. To change the choice later, open **Cookie settings** in the website footer. *** ## FAQ **Will a visitor's widget conversation return after a page reload?** Yes. The widget can restore the existing conversation so the visitor does not have to start again. **Does navigation always reload the website?** Not necessarily. Supported same-site navigation can move visitors between pages while keeping the widget open. Other destinations may still require a normal page load. **Which files can visitors upload to a Brand Agent?** The customer-facing Brand Agent upload flow supports images and PDFs. **What happens when a Free-plan crawl reaches 24 pages?** The crawl interface shows that the plan limit was reached and provides an upgrade path. You can also narrow the website source to stay within the limit. **Can visitors change their cookie preferences later?** Yes. They can reopen **Cookie settings** from the website footer. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Create a Brand Agent:** [/brand-agents/create-brand-agent](/brand-agents/create-brand-agent) * **Knowledge Bases:** [/shared-platform/knowledge-bases](/shared-platform/knowledge-bases) * **Deploy on a website:** [/deploy/custom-website](/deploy/custom-website) # Custom Skills, Smarter Template Setup & Brand Agent Sharing (June 18, 2026) Source: https://docs.sketricgen.ai/changelog/2026-06-18-custom-skills-template-setup-agent-sharing SketricGen update: custom skill installation, template-ready assets, faster connector setup, plan usage cards, Brand Agent sharing, and Code Interpreter access. This release makes advanced AI Workforce setups easier to install and reuse. Custom skills and template assets now arrive more reliably, connectors are matched more intelligently, and Brand Agents are easier to share. ## TL;DR * **Install custom skills:** Add packaged custom skills to your agents with validation and clear progress feedback. * **Templates bring their supporting assets:** Required skill files can be installed with a template and cleaned up when you switch templates. * **Smarter connector matching:** Imported templates find the appropriate connector provider and actions more reliably. * **Live template updates:** Template listings refresh more consistently and show more options at once. * **Plan usage at a glance:** Agent Space and Brand Agent pages show plan allowances and current usage. * **Share Brand Agents:** A new sharing experience makes it easier to send an agent to collaborators or testers. * **Code Interpreter available:** AI Workforce agents can use Code Interpreter without an additional feature toggle. *** ## What's new ### 1) Install and reuse custom skills You can now add custom skill packages to agents through a more complete installation flow: * **Packaged skill installation:** Install custom skills together with the files they need. * **Validation before use:** Skill packages are checked before they are made available to an agent. * **Visible progress:** Installation state is tracked so you can see when a skill is ready. * **Safer template changes:** Outdated skill artifacts are removed when you replace a template. **Why it matters:** You can reuse specialized agent behavior without manually rebuilding its instructions and supporting files every time. *** ### 2) Templates that set up more of the workflow Template installation now handles more of the configuration around the workflow itself: * **Supporting artifacts:** Templates can install the custom skill resources required by their agents. * **Cleaner template swaps:** Assets that belong to the previous template are pruned when they are no longer needed. * **Fresher listings:** Template data is refreshed when the page opens, reducing stale results. * **More discoverable options:** The initial template list shows up to 30 entries. **Why it matters:** Starting from a template requires less manual cleanup and produces a more complete, ready-to-test Agent Space. *** ### 3) Faster, more accurate connector setup Connector selection is more reliable when you install or import a workflow: * **Provider-aware matching:** Template tools are matched with the right available connector provider. * **Better action discovery:** Connector actions load from the current catalog when you need them. * **Safer cancellation:** Cancelling a template installation no longer disconnects an existing app connection. * **Faster initial loading:** Connector catalog details are loaded only when needed. **Why it matters:** Imported workflows are less likely to point at the wrong connector or disturb connections you already use. *** ### 4) Clearer limits and easier Brand Agent sharing Everyday account and Brand Agent management received several visible improvements: * **Plan usage cards:** Agent Space and Brand Agent pages show relevant limits and current usage. * **Brand Agent sharing:** Send an agent to teammates, reviewers, or testers from its page. * **Improved URL validation:** Brand Agent setup fields provide clearer feedback when a website address is incomplete or invalid. * **Smoother onboarding:** Guided tours better control when each step can continue. **Why it matters:** You can see available capacity before creating more agents and share work for review with fewer steps. *** ### 5) Code Interpreter for file and data work Code Interpreter is now available as an AI Workforce tool without a separate availability flag. * **Data analysis:** Let an agent calculate, transform, and inspect supported data files. * **File generation:** Use code-backed workflows to create useful output artifacts. * **Tool selection:** Add Code Interpreter from the tools available in the Agent Space editor. **Why it matters:** Agents can complete more calculation-heavy and file-based work inside the same workflow. *** ## Who this release is for * **AI Workforce builders** installing reusable custom skills. * **Template users** who want a more complete setup with less manual cleanup. * **Teams using multiple connector providers** in imported workflows. * **Brand Agent owners** sharing agents for review or testing. * **Workspace administrators** monitoring plan usage and limits. * **Data-focused teams** using Code Interpreter for analysis and file creation. ## How to use (quick steps) ### Install a custom skill 1. Open an agent in your **Agent Space**. 2. Open its **Skills** area and choose the custom skill installation option. 3. Select the skill package and wait for validation and installation to complete. 4. Test the agent to confirm the skill is available during the run. ### Start from a template 1. Browse [AI Workforce templates](https://www.sketricgen.ai/template). 2. Select a template and install it into your Agent Space. 3. Review the installed skills, supporting files, and connector actions. 4. Connect the required apps before running the workflow. ### Share a Brand Agent 1. Open the Brand Agent you want to share. 2. Use the available **Share** action. 3. Send the resulting link through your preferred channel. *** ## FAQ **What is a custom skill?** A custom skill is a reusable set of instructions and supporting resources that teaches an agent how to perform a specific kind of work. **Do templates install their required skill files automatically?** Supported templates can install their associated skill artifacts as part of setup. Always review the installed skills and test the workflow before using it with production data. **What happens to skill files when I switch templates?** Artifacts that belonged to the previous template can be removed automatically, reducing stale or conflicting resources. **Will cancelling template setup disconnect an app I already connected?** No. Cancelling template installation should preserve existing app connections. **Where can I see my current plan usage?** Plan usage cards appear on relevant Agent Space and Brand Agent pages. **Is Code Interpreter available to every plan?** Its presence no longer depends on a separate feature flag. Your current plan and the in-product tool picker remain the source of truth for availability and usage limits. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **AI Workforce Overview:** [/ai-workforce/overview](/ai-workforce/overview) * **Tools and Skills:** [/user-guides/tools](/user-guides/tools) * **Create a Brand Agent:** [/brand-agents/create-brand-agent](/brand-agents/create-brand-agent) # Brand Agent Templates & Smoother Mobile Editing (July 8, 2026) Source: https://docs.sketricgen.ai/changelog/2026-07-08-brand-agent-templates-mobile-editor-fixes SketricGen update: dedicated Brand Agent template setup, corrected template links, and smoother Agent Space editing on mobile devices. This smaller release makes Brand Agent templates open through the right setup experience and improves Agent Space navigation on mobile devices. * **Dedicated Brand Agent template flow:** Brand Agent templates now open the appropriate Brand Agent setup instead of being treated like a standard AI Workforce template. * **Correct template links:** Template cards and calls to action now send both signed-in and signed-out visitors to the intended destination. * **Better mobile editing:** Agent Space navigation and editor layouts behave more reliably on smaller screens. * **More consistent template metadata:** Brand Agent templates use the same public content and search metadata expected across the template catalog. **Why it matters:** You can choose a Brand Agent template with confidence and continue configuring agents more comfortably from a mobile device. ## Get Started * **Browse Templates:** [https://www.sketricgen.ai/template](https://www.sketricgen.ai/template) * **Create a Brand Agent:** [/brand-agents/create-brand-agent](/brand-agents/create-brand-agent) * **Agent Space Canvas:** [/ai-workforce/agent-spaces](/ai-workforce/agent-spaces) # Teamspaces, Shared Plans & Project-Scoped Connections (July 13, 2026) Source: https://docs.sketricgen.ai/changelog/2026-07-13-teamspaces-shared-plans-project-connections SketricGen update: Teamspace collaboration, shared plan limits, project-scoped connections, premium model guidance, and spreadsheet artifacts. Teamspaces are now the center of collaborative work in SketricGen. This release gives teams shared access, shared plan allowances, project-specific connections, and more capable file workspaces while strengthening separation between organizations. ## TL;DR * **Teamspaces for shared work:** Invite teammates, share access, and manage ownership in one collaborative workspace. * **Shared plan and usage:** Members use the Teamspace owner's plan allowances, with limits tracked across the Teamspace. * **Teammate playground access:** Authorized members can test shared agents without needing the creator to run every test. * **Project-scoped connections:** Connected apps, custom credentials, and skills stay with the relevant Teamspace project. * **Clearer model access:** Premium models display plan requirements before you select them. * **Spreadsheet artifacts:** Run workspaces can open XLS and XLSX files in a spreadsheet-friendly viewer. * **Smoother chat:** Streamed responses animate more naturally and traces remain hidden until the response is ready. *** ## What's new ### 1) Teamspaces built for collaboration Teamspaces now support more of the ownership and access controls teams need for shared AI work: * **Member access:** Invite teammates to work within a shared Teamspace. * **Shared agent testing:** Authorized members can open the playground and test Teamspace agents. * **Ownership transfer:** Move Teamspace ownership when responsibilities change. * **Mobile switching:** Change Teamspaces and projects through a layout designed for smaller screens. * **Stronger isolation:** Access checks have been tightened so agents, runs, and credentials remain within the correct Teamspace. **Why it matters:** Your team can build and test together without passing work through one account or mixing resources between organizations. *** ### 2) Shared plans, limits, and billing context Usage and subscription behavior now follow the active Teamspace: * **Owner plan inheritance:** Shared Teamspaces use the creator's current plan and entitlements. * **Teamspace-wide usage:** Seat, agent, and knowledge base allowances are tracked for the whole Teamspace. * **Relevant checkout:** Upgrade and billing actions apply to the Teamspace you currently have selected. * **Reliable rewards:** Eligible referral and reward credits are applied more safely. **Why it matters:** Members see one consistent set of allowances, and billing changes affect the workspace they intend to manage. *** ### 3) Connections and credentials that belong to the project App access now follows the Teamspace project instead of an individual member: * **Project-scoped connections:** Connector access is associated with the project where agents use it. * **Shared custom connections:** Bring-your-own-key and remote connections can be available to the appropriate Teamspace. * **Consistent skill installation:** Custom skills install against the Teamspace owner and remain usable by the shared project. * **Safer separation:** Moving between projects no longer implies that every credential should follow you. **Why it matters:** Teams can collaborate on an agent while keeping its app access and custom credentials attached to the correct project. *** ### 4) Clearer model availability The model picker now communicates plan access before a run begins: * **Premium model labels:** Models that require a higher plan appear locked when they are unavailable. * **Plan-aware selection:** The catalog reflects the entitlements of the active Teamspace. * **Fewer failed setups:** You can choose an available model while configuring the agent instead of discovering the restriction at runtime. **Why it matters:** Model capabilities and plan requirements are easier to understand while you build. *** ### 5) Better workspace files and chat feedback AI Workforce runs are more useful for document and spreadsheet work: * **Teamspace file storage:** Conversation artifacts are organized under the relevant Teamspace and project. * **Spreadsheet viewing:** Open supported XLS and XLSX artifacts in a spreadsheet-friendly viewer. * **Smoother streaming:** Assistant text appears with a more natural animation during a run. * **Cleaner progress state:** The trace button stays out of the way until streaming is complete. **Why it matters:** Teams can inspect spreadsheet results directly while run feedback stays focused and readable. *** ## Who this release is for * **Teams collaborating on Agent Spaces** across multiple members. * **Workspace owners** managing seats, agents, knowledge bases, and billing. * **Operations teams** sharing connectors and project-specific credentials. * **Builders using premium models** who need clear plan guidance. * **Analysts and operators** working with spreadsheet artifacts. * **Mobile users** switching between Teamspaces and projects. ## How to use (quick steps) ### Collaborate in a Teamspace 1. Open your Teamspace settings and invite the teammates who need access. 2. Select the appropriate project before creating or editing an agent. 3. Ask an authorized teammate to open the agent's playground and run a test. 4. Review Teamspace usage before adding more members or resources. ### Configure a project connection 1. Switch to the Teamspace and project that will use the app. 2. Open **Tools** or the agent's connector setup. 3. Connect the app or add the supported custom connection. 4. Test the agent from the same project to confirm access. ### Review a spreadsheet artifact 1. Run an AI Workforce task that creates or updates a spreadsheet. 2. Open the run workspace or artifact panel. 3. Select the XLS or XLSX file to inspect it in the spreadsheet viewer. *** ## FAQ **Whose plan does a shared Teamspace use?** A shared Teamspace uses the owner's plan and entitlements. Usage and resource limits are tracked across that Teamspace. **Can teammates test an agent in the playground?** Yes. Authorized Teamspace members can test shared agents without signing in as the original creator. **Do my connected apps automatically appear in every project?** Connections are project-scoped. Connect or configure the app in the Teamspace project where the agent will use it. **What happens when I transfer Teamspace ownership?** The new owner takes responsibility for the Teamspace. Review membership, plan, billing, and connected resources as part of the transfer. **Why is a model shown as locked?** Locked models require a plan available to the active Teamspace. Use an unlocked model or review the available plan options. **Can I preview Excel files created by a run?** Yes. The run workspace supports spreadsheet-friendly viewing for XLS and XLSX artifacts. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **AI Workforce Overview:** [/ai-workforce/overview](/ai-workforce/overview) * **Agent Spaces:** [/ai-workforce/agent-spaces](/ai-workforce/agent-spaces) * **Runs, Files, and Artifacts:** [/ai-workforce/runs-files-artifacts](/ai-workforce/runs-files-artifacts) * **Tools and Skills:** [/user-guides/tools](/user-guides/tools) # Affiliate Rewards, Managed Connectors & Stronger Teamspace Controls (July 20, 2026) Source: https://docs.sketricgen.ai/changelog/2026-07-20-affiliates-managed-connectors-teamspace-controls SketricGen update: affiliate commissions, managed connector credits, premium models, Trial plan controls, Teamspace deletion, and higher knowledge base limits. This release expands the ways you can grow with SketricGen and gives workspace owners clearer control over plans and resources. It introduces affiliate commissions, makes SketricGen-managed connectors broadly available, strengthens premium model and Trial plan handling, and raises knowledge base allowances. ## TL;DR * **Affiliate program:** Refer new customers, track eligible activity, and earn commissions. * **Managed connectors:** Use supported apps with SketricGen-managed access and transparent credit billing. * **Clear premium model access:** Premium models require an active paid subscription, with a direct Trial plan checkout path when appropriate. * **More Trial plan control:** Trial checkout and cancellation now use dedicated confirmation dialogs. * **Delete unused free Teamspaces:** Owners can remove eligible free Teamspaces and their associated project records. * **Higher knowledge base allowances:** Plan tiers now support up to 25 knowledge bases, with updated size limits for each knowledge base. * **More reliable email:** Verification and Teamspace invitation emails use improved production delivery. * **Fresher blog pages:** The blog index is less likely to show stale content or fail while new posts are published. *** ## What's new ### 1) Earn with the SketricGen affiliate program SketricGen now includes a fuller affiliate experience for partners and advocates: * **Referral codes:** Share your referral identity with prospective customers. * **Eligible signup tracking:** Attribute new accounts that join through your referral. * **Subscription benefits:** Track qualifying subscription activity connected to referred customers. * **Commission tracking:** Eligible affiliate commissions are recorded as subscription events are processed. **Why it matters:** Creators, consultants, and partners can recommend SketricGen through a structured program and receive credit for customers they introduce. *** ### 2) SketricGen-managed connectors and credit billing Supported connectors can now use credentials managed by SketricGen: * **Managed app access:** Connect supported apps without providing your own provider credential where managed access is offered. * **Credit-based usage:** Eligible managed connector actions consume SketricGen credits. * **Current app catalog:** The Apps experience uses the latest managed connector availability. * **Project-aware access:** Connector identity continues to follow the Teamspace project that uses it. **Why it matters:** More teams can connect useful business apps without separately sourcing and maintaining every underlying integration credential. *** ### 3) Premium models and Trial plan controls Model access and trial decisions now have clearer account checks: * **Paid model access:** Premium models require an active paid subscription. * **Plan-aware guidance:** Locked models indicate that a subscription is required. * **Trial checkout dialog:** Eligible users receive a focused path for starting a Trial plan. * **Trial cancellation confirmation:** Cancelling a trial requires a clear confirmation step. **Why it matters:** You know when a model requires a paid plan, and important trial decisions are less likely to happen accidentally. *** ### 4) Stronger Teamspace ownership controls Owners have more control over unused free workspaces: * **Owner-only deletion:** Only the Teamspace owner can delete an eligible free Teamspace or project. * **Associated cleanup:** Deletion removes the related workspace records rather than leaving an incomplete project behind. * **Free-tier allowances:** Newly created free Teamspaces receive the appropriate starting limits. * **Active billing protection:** Paid Teamspaces follow their subscription management flow instead of the free-workspace deletion path. **Why it matters:** Owners can clean up unused free workspaces while destructive account actions remain restricted to the right person. *** ### 5) Higher knowledge base capacity Knowledge base allowances have been expanded across plan tiers: * **Up to 25 knowledge bases:** Plan tiers now support a higher total number of knowledge bases. * **Updated per-knowledge-base limits:** Individual size allowances have been adjusted with the new plan structure. * **Teamspace-wide enforcement:** Usage is evaluated for the active Teamspace so members see consistent capacity. **Why it matters:** You can organize more sources by product, department, client, or use case without combining everything into a small number of knowledge bases. *** ### 6) More dependable emails and blog updates Several reliability improvements make important communication and public content more dependable: * **Account emails:** Production verification messages use improved delivery infrastructure. * **Teamspace invitations:** Invite emails use the same more reliable sending path. * **Fresher blog index:** Newly published posts and updates are less likely to be hidden by stale page data. * **Safer service credentials:** Internal credential handling was strengthened without changing how you connect supported apps in the interface. **Why it matters:** Important account messages are more likely to arrive, and public content remains available as it changes. *** ## Who this release is for * **Affiliates, creators, and consultants** recommending SketricGen. * **Teams using managed app connectors** without separate provider credentials. * **Builders evaluating premium models** and Trial plan options. * **Teamspace owners** cleaning up unused free projects. * **Knowledge-heavy teams** organizing sources across more knowledge bases. * **Administrators** relying on verification and invitation emails. ## How to use (quick steps) ### Use the affiliate program 1. Open the available affiliate area in your SketricGen account. 2. Copy your referral code or referral link. 3. Share it with an appropriate audience. 4. Return to the affiliate area to review eligible referrals and commissions. ### Connect a managed app 1. Open **Tools** or the **Apps** catalog in the correct Teamspace project. 2. Select an app that offers SketricGen-managed access. 3. Complete the connection flow and review its available actions. 4. Test the agent and monitor credit usage for managed connector actions. ### Delete an unused free Teamspace 1. Switch to the free Teamspace you own. 2. Open its settings and choose the deletion option. 3. Review the warning and confirm only after verifying that the Teamspace is no longer needed. ### Review model access 1. Open an agent's model picker. 2. Check whether the preferred model is available or marked as premium. 3. Select an available model, or follow the plan guidance to review subscription options. *** ## FAQ **Who can join the affiliate program?** Availability and eligibility are shown in the SketricGen affiliate experience. Review the current program terms before sharing a referral link. **Do managed connectors use my own API key?** Not when SketricGen-managed access is available and selected. Those actions use managed credentials and consume credits according to the current in-product pricing. **Can Trial users use premium models?** Premium models require an active paid subscription. The model picker and checkout dialogs show the options available to your account. **Can any Teamspace member delete the workspace?** No. Eligible free Teamspace deletion is restricted to the owner. **How many knowledge bases can I create?** Plan tiers now support up to 25 knowledge bases. Exact size and usage allowances are shown in the product because they can vary by plan. **Will deleting a free Teamspace affect a paid subscription?** The free-Teamspace deletion flow is separate from paid subscription management. Always confirm the selected Teamspace and review the warning before deleting it. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **AI Workforce Overview:** [/ai-workforce/overview](/ai-workforce/overview) * **Tools and Skills:** [/user-guides/tools](/user-guides/tools) * **Knowledge Bases:** [/shared-platform/knowledge-bases](/shared-platform/knowledge-bases) * **Manage Your Agent Space:** [/ai-workforce/agent-spaces](/ai-workforce/agent-spaces) # Shopify, Resilient Agent Runs & Easier Project Management (July 29, 2026) Source: https://docs.sketricgen.ai/changelog/2026-07-29-shopify-resilient-runs-project-billing SketricGen update: Shopify product recommendations, resilient agent runs, project controls, self-serve billing, and connector improvements. This release makes customer-facing agents more dependable and easier to manage. Shopify stores can connect Brand Agents to live product recommendations, active agent runs survive interruptions, and Teamspace owners gain clearer project and subscription controls. ## TL;DR * **Shopify integration:** Connect an eligible Brand Agent to Shopify and show live product recommendations inside the shopper widget. * **Runs survive interruptions:** Refreshing, switching tabs, or briefly losing your connection no longer starts a duplicate conversation. * **Faster conversation switching:** Recently opened conversations load more quickly while active responses continue streaming. * **Project controls:** Create projects from the switcher and safely delete projects you no longer need. * **Centralized API keys:** Personal and Teamspace API keys now live together in Teamspace settings. * **Self-serve subscriptions:** Cancel, resume, or open the billing portal from the pricing experience. * **More reliable connectors:** Managed app connections and WhatsApp setup report their status more consistently. *** ## What's new ### 1) Connect Brand Agents to Shopify The Shopify experience now links your store to an eligible SketricGen Brand Agent: * **Guided connection:** Sign in to SketricGen, choose an eligible Brand Agent, and complete the store-linking flow. * **Live product recommendations:** Shopper conversations can display product cards from the connected Shopify catalog. * **Widget customization:** Open the linked agent's widget settings to match the experience to your storefront. * **Safer store linking:** Connection attempts expire and cannot be reused, helping prevent a store from being linked with an old authorization. * **Reliable uninstall handling:** Removing the Shopify app cleanly disconnects the integration. **Why it matters:** Your Brand Agent can move from answering general questions to helping shoppers discover relevant products from your actual catalog. *** ### 2) Keep active runs through refreshes and reconnects SketricGen now restores in-progress agent runs after common browser and network interruptions: * **Automatic reconnection:** An active response can resume after a page refresh, tab switch, widget remount, or temporary connection loss. * **No duplicate conversation:** Reopening an in-progress run continues the existing conversation instead of creating a second one. * **Prompt restoration:** The active request appears again when you reconnect, so the resumed response retains its context. * **Multiple active runs:** If more than one run is active, SketricGen can restore each run independently. * **Cross-session cancellation:** Stopping a detached run reaches the active worker even when the request reconnects through a different server. * **Faster history navigation:** Cached conversation history makes switching between recent conversations feel more immediate. **Why it matters:** Long-running agents remain usable when real-world interruptions happen, without losing work or producing confusing duplicate threads. *** ### 3) Create and manage projects more easily Project and account controls are now easier to find and safer to use: * **Create from the switcher:** Start a new project without leaving the Teamspace and project selector. * **Complete project deletion:** Deleting a project removes the agents and knowledge bases that belong to it. * **Default-project protection:** SketricGen checks for a valid replacement before changing which project opens by default. * **Centralized API keys:** View or regenerate your personal API key from the same Teamspace tab used for administrator-managed keys. * **Editable account name:** Update your account name from your profile. * **Plan-aware model ordering:** Agent models are ordered to make options for your current subscription easier to understand. **Why it matters:** Teams can organize separate initiatives without hunting through unrelated settings, while destructive project changes follow a controlled cleanup process. *** ### 4) Manage subscriptions without contacting support The pricing experience now covers more of the subscription lifecycle: * **Cancel a subscription:** Schedule cancellation directly from the pricing page. * **Resume before cancellation:** Choose **Don't cancel** before the scheduled end date to keep the subscription active. * **Open the billing portal:** Manage supported billing details through the customer portal. * **Clear cancellation date:** See when a scheduled cancellation will take effect. * **Trial confirmation:** Review a dedicated confirmation dialog before cancelling a trial. * **Affiliate reliability:** Eligible commissions remain associated with invoices when a subscription is replaced. **Why it matters:** You can make routine billing decisions yourself and see their timing before they take effect. *** ### 5) More dependable connectors and onboarding guidance Several smaller improvements make setup and connected apps clearer: * **Accurate connection state:** An active managed connection now appears as connected when you add or reopen its tool node. * **Environment-aware availability:** Only managed connectors enabled for the current SketricGen environment are offered. * **WhatsApp setup reliability:** Registration uses the configured Meta application and handles setup failures more clearly. * **Cleaner channel settings:** WhatsApp, Telegram, and Instagram management dialogs focus on settings that are currently supported. * **Onboarding video:** Eligible users can reopen the onboarding video from the dashboard. * **Responsive sign-in window:** The authentication popup displays more consistently across screen sizes. **Why it matters:** Setup screens more accurately reflect what is connected and give new users a clearer path when they need help. ## Who this release is for * **Shopify merchants** who want a Brand Agent to recommend products from their live catalog. * **Teams running longer agent tasks** where a refresh or temporary disconnect should not lose progress. * **Teamspace owners** creating, deleting, and organizing projects. * **Account administrators** managing API keys and subscription status. * **Affiliate partners** tracking eligible subscription commissions. * **Teams using WhatsApp or managed connectors** that need dependable connection status. ## How to use (quick steps) ### Connect a Brand Agent to Shopify 1. Install or open the SketricGen app from your Shopify admin. 2. Sign in to the SketricGen account that owns the Brand Agent. 3. Select an eligible **Brand Agent** and complete the connection. 4. Open the agent's **Widget** settings to customize the storefront experience. 5. Test a product question and confirm that the response can display catalog cards. ### Resume an interrupted agent run 1. Start an agent run in the Playground or an embedded widget. 2. If the page refreshes or the connection drops, reopen the same conversation. 3. Let SketricGen reconnect to the active run; do not submit the request again. 4. Use **Stop** if you need to cancel the restored run. ### Create or delete a project 1. Open the **Teamspace and project switcher**. 2. Choose the option to create a project, then enter its name. 3. To remove a project, open its project settings and review the deletion warning. 4. Confirm only after checking the agents and knowledge bases that will also be removed. ### Manage a scheduled cancellation 1. Open the **Pricing** page from the Teamspace whose subscription you want to manage. 2. Choose the available cancellation or billing-portal action. 3. Review the effective date before confirming. 4. If the cancellation is still pending, choose **Don't cancel** to resume the subscription. *** ## FAQ **Will refreshing an active run create another conversation?** No. SketricGen attempts to reconnect to the existing run and continue from its preserved position. If the run has already finished, the completed conversation appears in history. **Does deleting a project also delete its agents and knowledge bases?** Yes. Project deletion includes the resources owned by that project. Review the confirmation carefully and export or move anything you still need before deleting it. **Which agents can I connect to Shopify?** The Shopify selector lists Brand Agents eligible for the connection. If an agent does not appear, confirm that it is a Brand Agent in the active Teamspace project. **Do Shopify product cards use my live catalog?** Yes. Recommendations in the connected shopper widget use products available through the store's catalog connection. **When does a scheduled subscription cancellation take effect?** The pricing page shows the scheduled cancellation date. You can resume the subscription before that date when the **Don't cancel** action is available. **Where did my personal API key move?** Personal and Teamspace-managed API keys are now grouped in the API keys tab of Teamspace settings. Your permissions determine which keys and actions you can see. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Shopify deployment guide:** [/deploy/shopify](/deploy/shopify) * **Agent Playground:** [/ai-workforce/playground](/ai-workforce/playground) * **Teamspace collaboration:** [/ai-workforce/team-collaboration](/ai-workforce/team-collaboration) * **Knowledge Bases:** [/shared-platform/knowledge-bases](/shared-platform/knowledge-bases) # Safer Widgets, Expanded Brand Agent APIs & Wix Improvements (August 6, 2026) Source: https://docs.sketricgen.ai/changelog/2026-08-06-secure-widgets-brand-agent-api-wix SketricGen update: safer embedded widgets, expanded Brand Agent APIs and MCP tools, stronger Wix linking, and more reliable AI-assisted building. This release strengthens how embedded agents authenticate visitors and expands the ways developers and AI clients can manage Brand Agents. Wix linking is safer, while Agent Build's Import, Improve, and Skill assistants are more responsive during longer requests. ## TL;DR * **Safer public widgets:** Embedded agents use short-lived, agent-specific authorization instead of exposing a reusable runtime key. * **Visitor continuity:** Returning widget visitors can continue their identity without letting the browser choose a trusted contact identifier. * **Private widgets preserved:** API-key-protected agents continue to enforce their existing access requirements. * **Editable Brand Agents by API:** Update Brand Agent details, models, knowledge bases, and widget behavior without replacing the workflow graph. * **Richer MCP tools:** AI clients can discover Brand Agent templates and knowledge bases, then use them during agent creation. * **Connector-aware automation:** Brand connectors can be configured through the Admin API and MCP workflows with tool-level validation. * **Stronger Wix linking:** Verified Wix site context reaches the widget while invalid connection data is rejected safely. * **More reliable AI assistance:** Import, Improve, and Skill operations can stream longer work without the previous short request limit. *** ## What's new ### 1) Safer embedded widgets and visitor sessions Public and private widgets now use a more narrowly scoped authorization model: * **Short-lived access:** Public widgets request temporary authorization for one agent instead of receiving a reusable runtime API key in the embed page. * **Automatic renewal:** Active widget sessions refresh their temporary authorization while a conversation is running. * **Signed visitor continuity:** SketricGen can recognize a returning anonymous visitor without trusting an identity supplied directly by the browser. * **Consistent fullscreen behavior:** Fullscreen and iframe widgets follow the same visitor identity rules. * **Private-agent protection:** Agents that require an API key keep that additional access check. * **Tighter agent permissions:** Import, instruction improvement, skill generation, lead submission, and other protected actions verify access to the target agent. **Why it matters:** You can embed agents publicly without placing a long-lived platform credential in customer-facing HTML, while private agents remain private. *** ### 2) Manage more Brand Agent settings through the API The Brand Agent management API now supports practical edits after creation: * **Agent details:** Update display names, agent names, instructions, the selected model, and attached knowledge bases. * **Widget configuration:** Change supported widget appearance and behavior settings independently from the agent's core details. * **Safe partial updates:** Send only the fields you want to change; the API preserves the remaining stored configuration. * **Brand Agent boundaries:** These operations cannot add or remove workflow nodes and reject agents that are not Brand Agents. * **Plan enforcement:** Model access and branding removal follow the Teamspace's active subscription entitlements. * **Clearer failures:** Invalid models, unavailable widget configuration, and unsupported agent types return explicit errors instead of silently overwriting data. **Why it matters:** Developers and AI operators can maintain Brand Agents programmatically while SketricGen preserves workflow structure and subscription rules. **For developers:** Use `PATCH /brand-agents/{agent_id}` for Brand Agent details and knowledge base assignment, and `PATCH /brand-agents/{agent_id}/widget-config` for widget settings. The retired standalone knowledge base attach and detach endpoints should no longer be used. *** ### 3) Build Brand Agents through richer MCP tools The SketricGen Model Context Protocol (MCP) server can now supply more of the context an AI client needs: * **Discover templates:** List available Brand Agent templates before choosing how to create an agent. * **Discover knowledge bases:** List accessible knowledge bases and select one during creation. * **Template-based creation:** Create a Brand Agent from a template identifier instead of recreating its setup manually. * **Knowledge-grounded creation:** Attach a selected knowledge base as part of the agent creation request. * **Connector setup:** Add supported Brand Agent connectors and grant only tools available for the selected app. * **Job visibility:** Creation-job results include the connectors being prepared for the agent. **Why it matters:** Claude, ChatGPT, and other MCP-compatible AI clients can understand available building blocks and assemble a more complete Brand Agent with fewer manual follow-up steps. *** ### 4) Stronger Wix site connections Wix integrations now pass verified site information into the connected agent experience: * **Site-aware widgets:** The widget receives the Wix site and instance context it needs for connected actions. * **Verified connection data:** SketricGen validates site information before treating it as the live Wix tool endpoint. * **Safe retries:** Invalid connection data is rejected without consuming the one-time link, so you can correct the issue and try again. * **Graceful version compatibility:** Older Wix app versions can still activate when optional signed context is absent. * **Regenerated embed bundle:** The production Wix embed code includes the latest connection behavior. **Why it matters:** Agents connected to Wix can act with the right site context, and failed setup attempts are less likely to leave the connection stuck. *** ### 5) Longer, more responsive AI-assisted building Three Agent Build assistants now use a dedicated streaming path: * **Import:** Watch progress while SketricGen converts an external API definition into agent-ready configuration. * **Improve:** Get more reliable instruction-improvement results without depending on a short page request. * **Skill:** Stream generated skill content as it is produced. * **Protected target selection:** The browser selects the feature, while SketricGen controls which internal assistant performs the work. * **Conversation reliability:** Additional history-refresh handling keeps widget transcripts consistent as conversations update. **Why it matters:** More complex imports and generations can take the time they need while still showing progress and enforcing access to the agent being edited. ## Who this release is for * **Teams embedding public agents** that want safer visitor authorization without managing exposed runtime credentials. * **Teams operating private agents** that rely on API-key-gated widget access. * **Developers using the Admin API** to update Brand Agents after creation. * **AI clients and agentic tools using MCP** to discover templates, knowledge bases, and connectors. * **Wix site owners** connecting a Brand Agent to a specific site. * **Builders using Import, Improve, or Skill** for more involved Agent Build tasks. ## How to use (quick steps) ### Update a Brand Agent through the API 1. Authenticate with credentials that can manage the target Brand Agent. 2. Send a partial update to the Brand Agent details or widget configuration endpoint. 3. Include only the fields you intend to change. 4. Handle entitlement or validation errors before retrying with a supported value. 5. Open the Playground or widget preview to verify the result. ### Create a Brand Agent with an MCP client 1. Connect your MCP-compatible AI client to the SketricGen MCP server. 2. Ask it to list available Brand Agent templates and knowledge bases. 3. Choose a template and, when relevant, a knowledge base. 4. Ask the client to create the Brand Agent with those identifiers. 5. Review the completed agent and test any requested connectors before publishing it. ### Connect a Wix site 1. Start the SketricGen connection flow from the Wix app. 2. Choose the Brand Agent you want to use on that site. 3. Complete the one-time linking step. 4. If verification fails, restart or retry the link instead of creating a separate connection. 5. Test the widget from the live Wix site so it receives the correct site context. ### Use an AI-assisted Agent Build action 1. Open the agent you are allowed to edit in **Agent Build**. 2. Start **Import**, **Improve**, or **Skill** from its relevant editor. 3. Keep the progress view open while the response streams. 4. Review the generated configuration before saving or publishing the agent. *** ## FAQ **Does a public widget still expose a SketricGen runtime API key?** No. Public widget pages request short-lived authorization scoped to the selected agent. The temporary credential is renewed as needed instead of embedding a long-lived runtime key in the page. **Did private widget authentication change?** Private agents still require their existing API-key-based access. The new temporary widget authorization does not make a private agent public. **Can the Brand Agent update API redesign my workflow graph?** No. It updates supported Brand Agent fields and widget settings, but it cannot add or remove workflow nodes. Use Agent Build for graph-level changes. **How should API clients attach a knowledge base now?** Include the desired knowledge base identifiers in the Brand Agent update. The older standalone attach and detach operations were removed because they did not reliably update the agent's actual knowledge tool configuration. **Will an API or MCP request bypass my plan limits?** No. Model selection, branding removal, and other paid capabilities are validated against the active Teamspace plan even when the change comes from an API or MCP client. **What happens if Wix cannot verify the site information?** SketricGen rejects invalid signed context and preserves the one-time link for a valid retry. If optional context is simply missing from an older app version, the base connection can continue without trusting unverified site data. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Public API guide:** [/dev-guide/public-api-usage](/dev-guide/public-api-usage) * **Structured Output tools:** [/dev-guide/structured-output-tools](/dev-guide/structured-output-tools) * **Wix deployment guide:** [/deploy/wix](/deploy/wix) * **Brand Agent connectors:** [/brand-agents/connectors-and-leads](/brand-agents/connectors-and-leads) * **Agent Build:** [/ai-workforce/agent-build](/ai-workforce/agent-build) # Collaborative Agent Space, Agent Build & Adaptive Onboarding (August 17, 2026) Source: https://docs.sketricgen.ai/changelog/2026-08-17-collaborative-agent-space-agent-build-onboarding SketricGen update: collaborative Agent Space, focused Agent Build, adaptive onboarding, transparent credits, and improved team workflows. This release introduces a clearer way to operate an AI workforce. **Agent Space** is now the shared project-level view of your agents and teammates, while **Agent Build** is the focused canvas for understanding and changing how one agent works. New users also get an adaptive onboarding conversation that uses their website and goals to recommend a useful starting team instead of presenting the same setup to everyone. ## TL;DR * **Collaborative Agent Space:** See every top-level agent in a project, understand its status, and spot teammates working with it. * **Shared live conversations:** Join a teammate's active Playground conversation and collaborate with the same agent in real time. * **Focused Agent Build:** Open one agent's internal workflow, edit its nodes, and move directly into testing. * **Structured Output and Human Input:** Define predictable response data and let an agent pause for a person's decision when needed. * **Adaptive onboarding:** Share your website, role, and goals to receive personalized agent recommendations and start a Brand Agent. * **More transparent credits:** Annual plan credits release monthly, eligible users can claim a daily allowance, and add-on purchases have clearer confirmation. * **Improved history and privacy:** Conversation visibility, pagination, private conversations, and retention are handled more consistently. * **Richer Teamspaces:** Switch projects more reliably, add project icons and profile photos, and see when an app connection is shared with a project. *** ## What's new ### 1) Operate your AI workforce from Agent Space Agent Space is the project-level canvas for seeing what your agents and teammates are doing: * **One project view:** See every top-level agent in the active project without opening each workflow. * **Operational filters:** Group agents by **Needs attention**, **Working**, **Scheduled**, or **Sleeping**, with a count for each state. * **Agent details:** Select an agent to inspect its role and start a quick conversation from the canvas. * **Draggable layout:** Arrange agent avatars in a layout that makes sense to you. * **Personal canvas memory:** Positions and zoom are remembered for each browser, member, and project. Moving an agent does not rearrange another teammate's canvas. * **Accessible navigation:** Use visible zoom controls or pinch gestures, and rely on reduced-motion behavior when enabled in your device preferences. Agent Space showing the project workforce, status filters, and selected agent details **Why it matters:** Instead of treating agents as isolated chatbots, your team gets an operational view of the workforce, including which agents are active and which need attention. *** ### 2) Collaborate with teammates in live agent conversations Agent Space shows live human presence around the agents your team is using: * **Live teammate presence:** A teammate appears beside an agent while they have that agent's shared Playground open. * **Join the same conversation:** Select the teammate presence indicator to enter the conversation they are using. * **Shared streaming response:** Participants in the shared conversation see the agent's response progress without waiting for separate sessions to refresh. * **Inline quick chat:** Start and stop an agent response directly from Agent Space when you need a fast interaction. * **Membership-aware access:** Presence and conversations are limited to people who can access the project and agent. * **No artificial activity:** Merely viewing Agent Space does not make you appear as actively chatting with an agent. **Why it matters:** Multiple team members can work with the same agent as a shared collaborator, reducing duplicate prompts and making the agent's current work visible to the project. *** ### 3) Understand and edit one agent in Agent Build Agent Build is the per-agent canvas that exposes how the selected agent works: * **Clear product boundary:** Agent Space manages the workforce; Agent Build manages one agent's internal workflow. * **Node-level editing:** Inspect agents, tools, skills, knowledge sources, and other workflow steps from the canvas. * **Unsaved-change protection:** If you leave a node with draft edits, SketricGen asks before discarding them. * **Contained configuration panels:** Long node settings remain inside the inspector instead of overflowing the editing area. * **Validation before use:** Skill and tool limits are checked before the workflow is accepted. * **Direct testing path:** Move from Agent Build into the Playground to test the current agent behavior. Agent Build canvas showing an agent's internal workflow and configuration panel **Why it matters:** You can see what gives an agent its behavior, change one component at a time, and test the result without treating the agent as a black box. *** ### 4) Add Structured Output and Human Input Agent workflows can now handle two important production patterns more explicitly: * **Structured Output:** Define the fields and data shape an agent should return when another system needs predictable output. * **Schema-based validation:** Structured Output uses a versioned JSON Schema contract so downstream tools and AI clients can understand the expected result. * **Human-in-the-Loop:** Let an agent pause and ask a person for input before continuing a sensitive or ambiguous step. * **Skippable questions:** Builders can configure Human Input so a user may skip the question when it is optional. * **Reliable resume behavior:** Answered or outdated questions do not reappear when a paused conversation resumes. * **Turn protection:** The composer remains unavailable while a run is still active, preventing a second message from conflicting with the current operation. **Why it matters:** Agents can produce machine-readable results when automation needs certainty and request human judgment when automation should not decide alone. *** ### 5) Start with adaptive, agent-driven onboarding The new onboarding experience adapts its recommendations to the person and business setting up SketricGen: * **Website context:** Enter a website with or without `https://` so SketricGen can understand the business and begin collecting relevant knowledge. * **Role and goal discovery:** Tell the onboarding agent what you do and which outcomes matter most. * **Personalized recommendations:** Receive marketing, growth, and other agent suggestions based on your website and stated needs. * **Brand Agent creation:** Review the Brand Agent decision inside the onboarding conversation while its setup starts in parallel. * **Progress persistence:** Resume onboarding without replaying steps you already completed. * **Safe navigation:** Website knowledge collection can continue across onboarding page transitions. * **Optional skip:** Skip onboarding when you already know how you want to build. Adaptive onboarding collecting role, goals, and website preferences **Why it matters:** A first-time user can begin with business outcomes and familiar website context instead of needing to understand agent architecture before seeing value. *** ### 6) Clearer credits, allowances, and add-ons Credit behavior is now easier to connect to the plan and purchases that provide it: * **Monthly release for annual plans:** Annual subscriptions receive one month's plan credits at a time on an automatic schedule instead of receiving the entire annual amount at once. * **Daily allowance:** When an eligible signed-in user's balance is too low to start a turn, they can claim a top-up once per rolling 24-hour period. * **No automatic resend:** Claiming the allowance does not silently submit the failed message again; you decide when to retry it. * **Public widget stability:** The daily allowance does not change how public widget visitors consume credits. * **Knowledge pack management:** Add, change, or remove knowledge pack quantities from an active subscription. * **Purchase confirmation:** Add-on and seat purchases show a confirmation step and clearer checkout progress. * **Current pricing:** Pricing data is refreshed from the billing system instead of depending on stale cached values. **Why it matters:** Teams can understand when credits become available, deliberately retry a blocked request, and review paid changes before checkout. *** ### 7) More consistent conversations, Teamspaces, and integrations Several platform improvements make shared work and connected systems easier to reason about: * **Conversation retention:** The history available to a Teamspace follows the retention included with its current plan. * **Private conversations:** Conversation visibility respects private-conversation rules in addition to project membership. * **Reliable long histories:** Cursor-based pagination loads additional messages without depending on unstable page positions. * **Project switching:** SketricGen reconciles the selected project when membership or subscription context changes. * **Project and profile images:** Upload a project icon and a personal profile picture to make shared spaces easier to recognize. * **Connection disclosure:** Before connecting an app, SketricGen explains that the connection is shared with the active project. * **Schedule safety:** Active schedule limits are enforced and unsafe trigger replacements fail visibly instead of silently replacing a working schedule. * **Shopify and Wix reliability:** Shopify product cards use the live Storefront catalog, while Wix connection requests can recover their required server secret correctly. **Why it matters:** Collaboration state, history, schedules, and shared connections behave more predictably as a Teamspace grows. ## Who this release is for * **New SketricGen users** who want recommendations without learning agent architecture first. * **Team leads and operators** who need a live view of agents and human activity across a project. * **Collaborative teams** that want to share an agent conversation instead of duplicating work in separate chats. * **Agent builders** configuring workflows, Structured Output, and Human Input. * **Annual subscribers and add-on buyers** who need transparent credit and billing behavior. * **Teams with sensitive conversations** that rely on privacy and plan-based history retention. * **Shopify and Wix users** running agents on connected websites. ## How to use (quick steps) ### Review and collaborate from Agent Space 1. Open **Agent Space** and choose the correct Teamspace project. 2. Use the status filters to find agents that are working, scheduled, sleeping, or need attention. 3. Select an agent to open its details or start a quick conversation. 4. If a teammate appears beside the agent, select their presence indicator to join the shared conversation. 5. Open **Agent Build** when you need to change the agent rather than chat with it. ### Add a Structured Output or Human Input step 1. Open the agent in **Agent Build**. 2. Add or select the relevant **Structured Output** or **Human Input** node. 3. Define the expected fields, or write the question a person should answer. 4. Decide whether the Human Input question may be skipped. 5. Save the workflow, then test both the normal and paused paths in the Playground. ### Complete adaptive onboarding 1. Enter your website on the SketricGen welcome screen. 2. Tell the onboarding agent about your role and desired outcome. 3. Review the recommended marketing, growth, or other specialist agents. 4. Confirm the Brand Agent path when it fits your use case. 5. Continue into Agent Space and test the created agents with a real task. ### Claim and use the daily allowance 1. Submit a Playground turn as an eligible signed-in user. 2. If the run cannot start because the balance is below the required threshold, review the allowance notice. 3. Claim the available daily top-up. 4. Resubmit the message when you are ready; SketricGen does not resend it automatically. 5. If no allowance is available, review plan or add-on options from the pricing experience. *** ## FAQ **What is the difference between Agent Space and Agent Build?** Agent Space is the shared project view of your workforce: agents, statuses, teammates, and quick conversations. Agent Build is the internal canvas for one agent: its instructions, nodes, tools, skills, knowledge, and execution structure. **Can two people chat with the same agent at the same time?** Yes. A teammate with the shared Playground open appears around that agent in Agent Space, and another authorized project member can join the same conversation and see the streamed response. **Does moving an agent on my Agent Space canvas move it for everyone?** No. Canvas position and viewport are stored for the current member, project, and browser. This lets each teammate arrange the workforce without changing another person's layout. **Does Structured Output guarantee that every agent response is JSON?** Structured Output defines the expected machine-readable result for the workflow step where you use it. Test the complete workflow and handle validation errors in the receiving system before relying on the output in production. **How does the daily allowance work?** An eligible signed-in user can claim it after a turn is rejected for insufficient starting credits, at most once per rolling 24 hours. It does not automatically resend the rejected message and does not change public-widget behavior. **Are app connections private to the person who creates them?** Project connections are shared with the active project. The connection flow now states this explicitly so you can confirm the correct project before authorizing the app. **What happened to MAX?** The separate MAX dashboard experience was removed. Agent Space, Agent Build, adaptive onboarding, and the Playground now provide the primary paths for discovering, operating, improving, and testing agents. *** ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Agent Space:** [/ai-workforce/agent-space](/ai-workforce/agent-space) * **Agent Build:** [/ai-workforce/agent-build](/ai-workforce/agent-build) * **Team collaboration:** [/ai-workforce/team-collaboration](/ai-workforce/team-collaboration) * **Adaptive onboarding:** [/getting-started/adaptive-onboarding](/getting-started/adaptive-onboarding) * **Structured Output:** [/dev-guide/structured-output-tools](/dev-guide/structured-output-tools) * **Agent Playground:** [/ai-workforce/playground](/ai-workforce/playground) # Workspace files, collaboration, safer embeds, and scoped API keys (September 9, 2026) Source: https://docs.sketricgen.ai/changelog/2026-09-09-workspace-files-collaboration-scoped-api-keys SketricGen update: manage workspace files, inspect agent activity and run costs, show web sources, restrict website embeds, create scoped API keys, and build on mobile. This release gives AI Workforce a clearer home for files, activity, and collaboration. It also makes web search evidence and run costs easier to inspect, adds better controls for embedded Brand Agents, and simplifies API key creation. ## TL;DR * **Scoped workspace files:** Browse Project, Agent, and Conversation files from one Files experience, then create, upload, edit, rename, or delete files when your permissions allow it. * **More visible agent work:** Agent Space shows creator and updater details, recent activity, teammates, active runs, and requests that need attention. * **Evidence and cost in traces:** Web search sources persist in conversations and traces, while run costs are split into model, managed app, web search, and code-execution rows. * **More useful human input:** Human Input questions can accept multiple choices, including a custom "Something else" response. * **Safer website embeds:** Restrict a public Brand Agent to approved domains, lazy-load the widget iframe, and use refreshed mobile and error states. * **Unified API keys:** Create one key for runtime access, administration, or both, with project and agent restrictions. * **A clearer first-agent path:** The Brand Agent wizard now includes a live preview, and onboarding saves preferences with retry support. * **Responsive product surfaces:** Agent Space, Playground, settings, and public pages work better on smaller screens. ## What's new ### 1) Browse and manage workspace files The new **Files** experience brings reusable and run-specific files into one place: * **Three storage scopes:** Switch between **Project**, **Agent**, and **Conversation** files. * **Project files:** Share reference material with every agent and conversation in the selected Project. * **Agent files:** Keep reusable files available across an agent's conversations. * **Conversation files:** Keep files private to one conversation. They are removed 30 days after the last write. * **Folder navigation:** Open nested folders with breadcrumbs and review file size and modified time. * **File management:** Create text files, upload files into the current folder, save edits, rename files, and delete files when your access allows it. * **Type-aware viewing:** Edit supported text files in the browser, preview images and PDFs, and download file types that cannot be previewed. Text files larger than 1 MB stay view-only in the browser. * **Safer edits:** Save and rename operations detect conflicts when a file changed after you opened it instead of silently overwriting newer work. The Files experience is available from the AI Workforce workspace and is also connected to the files used by runs and schedules. See [Playground](/ai-workforce/playground) and [Scheduled Runs, Files, and Artifacts](/ai-workforce/runs-files-artifacts) for the file scopes that agents can use. **Why it matters:** You can find the right file without opening a particular conversation, keep durable agent context separate from temporary run output, and make small text changes without leaving SketricGen. ### 2) See agent ownership and recent activity in Agent Space Agent Space now gives teams more context around the agents they share: * **Created by:** Agent cards show the person who created an agent when that information is available. * **Last updated by:** Cards and details show who last changed an agent when a user made the change. * **Recent activity:** Administrators can review recent agent creation, edit, and delete events from the details panel. * **Live collaboration:** Agent Space continues to show teammates, active runs, and requests waiting for human input. * **More resilient states:** Loading and retry actions make it clearer when live status is delayed or unavailable. * **Permission-aware activity:** Presence and activity do not grant access to an agent, Project, or conversation. See [Agent Space](/ai-workforce/agent-space) and [Multiplayer AI: Collaborate With Agents in Real Time](/ai-workforce/team-collaboration) for the current collaboration model. **Why it matters:** A team can tell who changed an agent, what happened recently, and which work needs attention without opening every conversation. ### 3) Show web search sources and support multi-choice human input Agent conversations and trace views now expose more of the evidence and decisions behind a run: * **Clickable web sources:** Web search sources appear below the relevant response and in the trace, with the source hostname and link. * **Persisted citations:** Sources remain available when a saved conversation or trace is reopened. * **Correct attribution:** Sources found by a delegated agent stay with that activity instead of being presented as evidence for an unrelated step. * **Search disclosure:** The trace can show that a search ran even when the provider returned no source links. * **Multi-select questions:** Human Input can ask users to select all applicable options rather than choosing only one. * **Additive custom answers:** A user can select normal choices and add a typed "Something else" response when the question supports it. * **Single-select compatibility:** Existing single-choice Human Input questions keep their current behavior. * **Shared Playground updates:** Teammate answers, paused runs, and streamed responses synchronize more reliably in shared conversations. For background on Human Input and run activity, see [Playground](/ai-workforce/playground) and [AI Workforce tools](/user-guides/tools). **Why it matters:** People can check where a time-sensitive answer came from and give an agent a more complete decision when one choice is not enough. ### 4) Understand where run credits went The trace details view now breaks down a run's total credit usage: * **Model calls:** Credits used by the agent's own model work. * **Managed apps:** Credits attributed to each SketricGen-managed connector used during the run. * **Web search:** Search activity is visible as its own row for clarity. It remains part of the model charge and is not added twice. * **Code execution:** Credits and run counts for code execution appear separately when the capability is used. * **Older traces:** Runs created before per-app detail was available still reconcile through a connected-app remainder row when needed. * **Large traces:** Traces stored outside the main record now load back into the trace view after access and privacy checks, so long runs do not appear blank. Open a run's trace from the Playground or Agent Build to review its activity, evidence, and credit usage before changing the workflow. **Why it matters:** The headline total now has an explanation. You can see whether a run spent credits on model work, a managed connector, web search, or code execution. ### 5) Control where a Brand Agent widget can run Public Brand Agent embeds now support an allowed-domain list: * **Approved hosts:** Add up to 25 hostnames or website URLs where the widget may run. * **Subdomain wildcards:** Use a pattern such as `*.example.com` to allow the site's subdomains. * **Clear default:** With no domains saved, a public widget can be embedded on any website. * **Embed behavior:** A restricted widget uses the allowed-domain check, and the fullscreen iframe option is unavailable while restrictions are active. * **Lazy loading:** The widget iframe can load when it is needed instead of competing with the initial page load. The first visitor message waits until the widget is ready. * **Refreshed integrations:** Generated embed snippets, mobile behavior, and widget error messages have been updated for the current embed bundle. To configure this, open a public Brand Agent's **Publish** or **Embed** settings, add the domains under **Website access**, and select **Save**. See [Deploy to a custom website](/deploy/custom-website) for the embed options. **Why it matters:** You can keep an embedded customer-facing agent tied to the websites you control while reducing the work it does before a visitor opens the chat. ### 6) Create one API key with exact access Teamspace settings now use one creation flow for runtime and administrative credentials: * **Choose capabilities:** Give a key **Run agents**, **Manage resources**, or both. * **Project scope:** Apply runtime access to all accessible Projects or a selected set of Projects. * **Agent restrictions:** Limit runtime access to selected agents inside the chosen Project scope. * **Administrative role:** Choose Viewer, Editor, or Admin permissions when the key manages Teamspace resources. * **Explicit administrative scope:** An administrative key can target one Project or the whole Teamspace according to its configured scope. * **Safer expiry:** Keys with administrative access use an expiry limit. Runtime-only keys can use the available runtime expiry options. * **Secret handling:** The new secret is shown once after creation, so save it before closing the confirmation step. Use [Teamspace Admin API](/dev-guide/admin-api) for management automation and the [Node.js SDK](/dev-guide/node-sdk) or [Python SDK](/dev-guide/python-sdk) for runtime integrations. **Why it matters:** One credential can match the job it performs. A widget or SDK does not need broad administrative access, and an automation key can be limited to the Projects and agents it actually needs. ### 7) Start faster and work comfortably on smaller screens The first-agent and product UI paths received a broad usability pass: * **Live Brand Agent preview:** The creation wizard shows a current preview while you set the website, sources, brand details, and widget options. * **Clearer onboarding:** The preference flow separates work role and AI-agent experience, saves each choice, and offers **Try again** when saving or loading fails. * **Better first-agent handoff:** Onboarding can recommend a Brand Agent and open its setup separately while the onboarding conversation remains available. * **Mobile access:** Protected product routes no longer stop at a desktop-only gate. Navigation, Teamspace switching, tabs, dialogs, Agent Space, and run views adapt to smaller screens. * **Reliable loading states:** Agent Space and run views show retry actions when a request stalls instead of leaving an empty panel. * **Public page metadata:** Landing pages, blog content, and templates use more reliable crawlable markup and structured data. Start with [Adaptive Onboarding](/getting-started/adaptive-onboarding), or skip it and open [Agent Space](/ai-workforce/agent-space) when you already know which workflow to build. **Why it matters:** New users can move from context to a useful first agent with less guesswork, and existing teams can manage their work from more devices. ### 8) Review the current schedule path Schedule setup continues through the Agent Build **Input** node and its **Scheduled Trigger** configuration. The legacy Pipedream trigger routes and scheduled-webhook service are no longer part of the platform. If an older Pipedream schedule still uses a callback URL containing `/schedule_webhook/`, migrate it to the current Scheduled Trigger flow or retire it before relying on that schedule. See [Scheduled Runs, Files, and Artifacts](/ai-workforce/runs-files-artifacts) for the current schedule setup. ## Who this release is for * **AI Workforce teams** managing shared Project, Agent, and Conversation files. * **Administrators** who need an activity trail and tightly scoped automation keys. * **Agent builders** inspecting web sources, credit usage, and multi-step human decisions. * **Brand Agent owners** embedding agents on websites they control. * **New users** starting with adaptive onboarding and a live Brand Agent preview. * **Mobile users** who need to build, test, and manage agents from smaller screens. ## How to use (quick steps) ### Manage workspace files 1. Open **Files** in AI Workforce. 2. Choose **Project**, **Agent**, or **Conversation** storage. 3. Select an agent or conversation when the chosen scope requires one. 4. Open folders and select a file to preview or edit it. 5. Use **New file**, **Upload**, **Rename**, **Delete**, or **Save** when your permissions allow the action. ### Review an agent's activity and trace costs 1. Open **Agent Space** for the selected Teamspace Project. 2. Select an agent to view its status, collaborators, activity, and pending input. 3. Open the agent's Playground or a completed run. 4. Open the trace and review sources, tool calls, credit rows, and generated files. ### Restrict a website embed 1. Open the public Brand Agent's **Publish** or **Embed** settings. 2. Find **Website access**. 3. Add a hostname or URL, or use a wildcard such as `*.example.com`. 4. Remove any domain that should no longer load the widget. 5. Select **Save**, copy the current embed snippet, and test it on an allowed site. ### Create a scoped API key 1. Open **Teamspace settings** and select **API keys**. 2. Select **Create key** and give it a name that identifies its use. 3. Choose **Run agents**, **Manage resources**, or both. 4. Set the Project and agent restrictions that the key needs. 5. Choose the administrative role and expiry when applicable. 6. Copy the secret from the confirmation step and store it securely. ## FAQ **Are Project files available to every agent?** Yes. Project files are shared with agents and conversations in the selected Project. Agent files remain scoped to one agent, and Conversation files remain scoped to one conversation. **Who can change a workspace file?** File actions depend on the selected scope and your Project permissions. If you do not have the required agent-edit access, the management actions are hidden. **Does web search add a second charge in the trace?** No. The trace shows web search as a separate line so you can see the activity, but that amount remains part of the model charge and is not added twice. **Can Human Input accept a normal choice and a custom answer together?** Yes, when the question enables multi-select and includes the "Something else" option. The custom text is submitted with the selected choices. **Can one API key run agents and manage resources?** Yes. Select both capabilities during creation, then set the shared Project scope and the appropriate administrative role. Keep runtime-only keys separate when a system only needs to run an agent. **Why is the fullscreen iframe option unavailable?** Fullscreen iframe embeds are unavailable while Website access restrictions are saved. Use a chat-bubble or floating-pill embed, or remove the domain restriction if a fullscreen iframe is required. **What should I do with an old Pipedream schedule?** Check its callback URL. If it contains `/schedule_webhook/`, migrate the schedule to the current Agent Build **Input** node and **Scheduled Trigger** flow, or retire it. ## Get Started * **Try SketricGen:** [https://www.sketricgen.ai/](https://www.sketricgen.ai/) * **Adaptive Onboarding:** [/getting-started/adaptive-onboarding](/getting-started/adaptive-onboarding) * **Agent Space:** [/ai-workforce/agent-space](/ai-workforce/agent-space) * **Playground:** [/ai-workforce/playground](/ai-workforce/playground) * **Runs, Files, and Artifacts:** [/ai-workforce/runs-files-artifacts](/ai-workforce/runs-files-artifacts) * **Team collaboration:** [/ai-workforce/team-collaboration](/ai-workforce/team-collaboration) * **Admin API:** [/dev-guide/admin-api](/dev-guide/admin-api) * **Custom website embed:** [/deploy/custom-website](/deploy/custom-website) # SketricGen v1.0.0 Source: https://docs.sketricgen.ai/changelog/v1.0.0 SketricGen v1.0.0 release notes — initial platform launch with AI agent builder, knowledge base, web search, and chat widget deployment. **Release Date:** *12 May 2025* SketricGen v1.0 was the first public release of the platform and introduced the foundation of the product’s no-code agent builder. ### Overview SketricGen v1.0 was built as a **single-agent system** focused on enabling anyone to design, test, and embed an intelligent assistant with their own data. ### Key Features * **Single Agent Creation** Users could create one agent at a time and define its role and behavior through a simple interface. * **Knowledge Base Integration** Each agent could connect to **one knowledge base**, which could include: * Uploaded files (PDF, DOCX, TXT, MD, PPTX, etc.) * Crawled websites * Q\&A pairs * Raw text inputs * **Playground Testing** A built-in playground allowed users to interact with their agent in real time and refine responses. * **Deployment Options** Agents could be embedded or published using: * **iFrame** or custom JavaScript widget snippets * Direct integrations with **Wix**, **WordPress**, **Squarespace**, **Shopify**, **Slack**, **WhatsApp**, and **Zapier** * **Conversations Log** A dedicated Conversations tab displayed all past interactions between users and the agent. * **Insights Dashboard** Users could view analytics such as usage statistics and conversational trends for performance monitoring. ### Summary Version 1 established the core functionality of SketricGen: a no-code environment to create data-driven AI agents that could learn from user-provided knowledge and be deployed anywhere. # SketricGen v2.0.0 Source: https://docs.sketricgen.ai/changelog/v2.0.0 SketricGen v2.0.0 release notes — major platform update with multi-agent orchestration, AgentSpace, app marketplace, and deployment integrations. **Release Date:** *08 October 2025* SketricGen v2.0.0 marks the platform’s most significant evolution to date — transforming it from a **single-agent builder** into a **multi-agent orchestration and deployment platform**. This release introduces **AgentSpace**, a visual workspace that lets users design complex workflows of interconnected agents, tools, and triggers — all within a no-code environment. ### Overview Version 2.0 reimagines how users build and manage AI agents. Instead of one isolated assistant, users can now design **entire agentic systems** — where multiple agents collaborate, exchange data, and perform specialized tasks through dynamic orchestration. ### New Core Features ### 1. AgentSpace (Canvas) A new visual workspace where users can: * Create multi-agent workflows with connected nodes. * Add **Agents**, **Tools**, **Triggers**, and **Handoffs** using a drag-and-drop interface. * Manage versions, promote drafts to production, and test directly from the AgentSpace. * Import or export configurations as JSON for version control and reuse. ### 2. Agents & Orchestration * Agents can now collaborate using **handoffs** or operate as **Agent-as-Tool** within other agents’ workflows. * Introduced **AI-routed orchestration**, where agents decide when to pass control based on their instructions and input context. * Added support for **Structured Inputs and Outputs** (typed JSON) to standardize communication between agents. * Users can define schemas directly in the UI — no coding required. ### 3. Tool Nodes & Extended Capabilities Agents can now call external services or perform computations via **Tool Nodes**, which include: * **Sketric App Marketplace:** Connect 2000+ apps (e.g., Gmail, Slack, Google Drive) through OAuth or API keys. * **File Search:** Build knowledge bases using uploaded files, crawled websites, raw text, or Q\&A pairs. * **Web Search:** Allow agents to retrieve live information from the web. * **Code Interpreter:** Run Python scripts for computation, file generation, or analysis. * **API Request:** Configure custom API endpoints with name, method, URL, and headers. * **Custom MCP:** Connect to Model Context Protocol servers for advanced interoperability. ### 4. Conversations & Traces * The **Conversations tab** now provides full trace visibility — users can inspect every message, agent decision, and tool call associated with each conversation. * The **Trace Explorer** remains integrated with AgentSpace for in-depth debugging at the workflow level. * Traces display agent interactions, structured inputs/outputs, credit usage, and tool performance in real time. ### 5. Insights Dashboard The **Insights tab** continues to track and visualize usage analytics for agents and workflows. It includes: * Conversation counts and traffic metrics * Credit consumption trends * Agent performance summaries and improvement patterns ### 6. Deployment & Integrations Expanded deployment options now allow users to publish multi-agent workflows directly to: * Websites (via JavaScript or iframe) * Third-party platforms such as **Slack**, **WhatsApp**, **Zapier**, **Wix**, **WordPress**, **Squarespace**, and **Shopify** * Public APIs using the new **/run-workflow** endpoint for real-time orchestration ### Additional Enhancements * **AI-powered Prompt Assistant:** “Improve your Instructions with AI” automatically refines agent prompts using best practices, with optional Edit Mode for manual adjustments. * **Configuration Management:** Users can now download, import, and duplicate full workflow setups as JSON configurations. * **Manage Tools Sidebar:** A new control panel to view, edit, and delete connected knowledge bases or third-party app integrations. * **UI/UX Enhancements:** Added zoom, pan, auto-layout, undo/redo, and workflow fit view options for improved design experience. ### Summary SketricGen v2.0.0 represents the transition from a simple AI assistant builder to a **full-featured agentic automation platform**. With multi-agent orchestration, tool integration, structured data handling, detailed traces, and enhanced analytics — users can now design intelligent, scalable systems that move beyond single-agent interaction into full workflow automation. # SketricGen v3.0.0 Source: https://docs.sketricgen.ai/changelog/v3.0.0 SketricGen v3.0.0 introduces Brand Agents for website chatbots and lead capture, plus AI Workforce for automation, files, skills, connectors, and models. **Release Date:** *25 May 2026* SketricGen v3.0.0 is a major release focused on making agents more practical, more powerful, and easier to deploy in the right place. ## TL;DR SketricGen now has two focused product surfaces: * **Brand Agents:** customer-facing website chatbots for leads, support, guided navigation, and public conversations. * **AI Workforce:** internal agents for workflow automation that can plan, call sub agents, use tools, work with files, create artifacts, and complete longer tasks. The release also adds a stronger Deep Agents runtime, 2,000+ app connectors, a searchable skill marketplace, richer traces, file workspaces, lead capture, and broader model choice. ## Core Features ### Brand Agents * Create a Brand Agent from a website URL or manual setup. * Extract Brand DNA from the business, website, and approved knowledge. * Structure website content into Knowledge Bases. * Add customer-facing tools such as lead capture, page navigation, web search, and selected app connectors. * Publish through a widget or fullscreen customer-facing chat experience. * Review conversations and captured leads after launch. ### AI Workforce * Build internal Agent Spaces around one main agent and focused sub agents. * Run longer tasks that can plan, use tools, call connectors, and create artifacts. * Use files and run workspaces during execution. * Start from AI Workforce templates and install reusable skills from the marketplace. * Search skills by task, review source and reputation signals, and install them directly onto an agent. * Inspect traces, tool calls, sub-agent calls, errors, latency, and credit usage. ### Connectors, Skills, and Models * Connect AI Workforce agents to broad app coverage through 2,000+ SketricGen connectors, API tools, and the Sketric connector marketplace. * Install marketplace skills so agents can use optimized task playbooks instead of starting from blank instructions. * Use curated connector access for Brand Agents so customer-facing agents only receive approved capabilities. * Choose from a broader model catalog instead of locking every agent to one model provider. ## Summary Version 3 moves SketricGen from a general multi-agent builder into a practical agent platform with separate paths for customer-facing agents and internal AI workforces. For the detailed customer-facing update, see [Practical Agents That Do Real Work](/changelog/2026-05-25-practical-agents-that-do-real-work). # Deploy an AI Chatbot to a Custom Website Source: https://docs.sketricgen.ai/deploy/custom-website Embed a SketricGen Brand Agent as an AI chatbot for your website with a floating chat widget, inline iframe, lead capture, and public deployment. Deploy a Brand Agent to your website when you want visitors to ask questions, get product guidance, or submit lead details. This creates a public website chat widget backed by your Brand Agent knowledge and approved tools. Use this guide for copy-paste website embeds. If you are building a custom product integration or calling your agent from code, use the [Public API](/dev-guide/public-api-usage) instead. ## Before You Start You need: * A created Brand Agent * A tested customer-facing conversation flow * Public/published visibility enabled * Widget styling reviewed The widget must be public because client-side website code cannot safely include a private API key. ## Website access restrictions By default, a public Brand Agent has no domain restrictions and its widget can be embedded on any website. Add an allowlist when the widget should run only on sites you control: 1. Open your Brand Agent's **Publish** or **Embed** settings. 2. Under **Website access**, enter a hostname such as `example.com` or a website URL, then select **Add**. 3. Add up to 25 domains. Use `*.example.com` to allow `example.com` and its subdomains. 4. Remove any domain that should no longer load the widget, then select **Save**. The allowlist checks the embedding page's HTTP or HTTPS hostname. Paths are not stored as separate rules, and a hostname that is not listed cannot load the restricted widget. If the widget is missing after you save the list, check that the page hostname matches an entry, including the correct subdomain or wildcard, and confirm that the Brand Agent is still public and published. Saving one or more allowed domains disables the **Fullscreen iframe** option. Use the chat bubble or floating pill embed while the widget is restricted, or remove all saved domains before using a fullscreen iframe. ## Floating Chat Bubble Use the floating chat bubble when you want the agent available across your site as a website chatbot. 1. Open your Brand Agent. 2. Test the agent with realistic visitor questions. 3. Open **Publish**. 4. Choose the chat bubble/widget option. 5. Copy the embed snippet. 6. Paste the snippet into your website template, layout, or tag manager. Example: ```html theme={null} ``` For multi-page websites, add the script to your shared layout so the widget appears consistently. ## Identify Known Users with Contact ID If your website or product already has logged-in users, pass your internal user or contact identifier with `data-contact-id`. SketricGen stores this as the conversation `contact_id`, so conversations from the same person are grouped under the same contact in SketricGen. Use a stable internal ID, UUID, or hashed value. Avoid sending emails, phone numbers, names, or other directly identifiable personal data as the contact ID. If you do not know what internal user ID to use, skip this field. The widget still works without it. Example for products with logged-in users: ```html theme={null} ``` If the visitor is anonymous, omit `data-contact-id`. The widget will use its browser-stored anonymous contact behavior instead. Developers can see the full identity model, including API, iframe, and fullscreen examples, in [Public API: Contact ID and User Grouping](/dev-guide/public-api-usage#contact-id-and-user-grouping). ## Inline iFrame Use an iframe when you want the agent embedded directly inside a page section. The **Fullscreen iframe** option is unavailable while one or more Website access domains are saved. Use a chat bubble or floating pill for a restricted embed, or remove the restrictions before generating a fullscreen iframe snippet. 1. Open your Brand Agent. 2. Open **Publish**. 3. Choose the direct embed or iframe option. 4. Copy the iframe code. 5. Paste it into the page where the chat should appear. Example: ```html theme={null} ``` To group iframe conversations for a known user, add the `contactId` query parameter: ```html theme={null} ``` ## Test After Deploying After adding the widget: 1. Open the live page. 2. Confirm the widget or iframe appears. 3. Ask a question from your Brand Agent's Knowledge Base. 4. If you pass `data-contact-id` or `contactId`, confirm the conversation shows under the expected contact in SketricGen. 5. Test lead capture if enabled. 6. Check the conversation in SketricGen. If the widget does not appear, confirm the Brand Agent is published/public and that the `data-agent-id` or iframe URL uses the correct agent ID. # Deploy to Framer Source: https://docs.sketricgen.ai/deploy/framer Embed a SketricGen Brand Agent as an AI chatbot for your Framer website with the SketricGen chat widget. This guide walks you through integrating your SketricGen Brand Agent with your Framer website using the chat widget embed code. Use it when you want a website chatbot for visitor questions, lead capture, product guidance, or support intake. ## Step 1: Create Your SketricGen Account To get started with the integration, [sign in to your SketricGen account](https://www.sketricgen.ai). If you don't have an account yet, you can [create one for free](https://www.sketricgen.ai/signup). Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. If you need assistance with setting up a working SketricGen Brand Agent, refer to our [step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart) for creating and deploying your Brand Agent. ## Step 2: Copy the SketricGen Brand Agent Embed Code After configuring your Brand Agent, go to the Brand Agent dashboard and click **Publish** to copy the embed code. **Important:** Ensure your Brand Agent is set to **Public** in order to add it to your Framer website. Screenshot showing SketricGen dashboard with Publish button and embed code copy option for the Brand Agent ## Step 3: Add Code to Your Framer Site ### Access Site Settings 1. Log in to your Framer account and open the site where you want the SketricGen Brand Agent widget to appear 2. Navigate to **Site Settings** Screenshot showing Framer Site Settings menu where Custom Code can be added ### Insert Custom Code 1. Go to the **Custom Code** section in site settings 2. Paste the embed code you copied in Step 2 3. Click **Save** Screenshot showing Framer Custom Code section with embed script pasted for SketricGen widget ### Publish Your Changes Publish your site to make the Brand Agent widget live and visible to visitors. ## Step 4: Test Your Integration After publishing, visit your Framer website to test the chat widget functionality. Screenshot showing SketricGen chat widget embedded and visible on a Framer website **Success:** You have successfully integrated the SketricGen Brand Agent widget. Visitors can now chat with your Brand Agent and get tailored responses. # Deploy to Instagram Source: https://docs.sketricgen.ai/deploy/instagram Connect your SketricGen Brand Agent to Instagram DMs. ## **Step 1: Sign Into SketricGen and Set Up Your Brand Agent** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don’t have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Connect Instagram Account to Your SketricGen Agent** Go to the **Publish** section of your Brand Agent and press **Connect** on the Instagram integration card. Screenshot showing the Instagram integration panel inside the Publish sidebar with a Connect button ## **Step 3: Grant Permissions to Your Instagram Account** Log in to your Instagram account when prompted and click **Allow** to grant the requested permissions to SketricGen. Screenshot showing the Instagram OAuth permissions prompt requesting access for SketricGen ## **Step 4: Confirm a Successful Connection** Once your Instagram account is successfully connected, the **Connect** button will switch to **Manage** inside the Publish sidebar. Click **Manage** to view or update the connected account details. Screenshot showing the Instagram integration card after connection with the Manage button and account details visible **Success:** Your Instagram account is now linked to **SketricGen**. Your Brand Agent can now handle Instagram conversations automatically. **Note:** Instagram has a 1000 character limit per message. If your Brand Agent produces responses longer than this limit, the message will be automatically split into multiple chunks and sent sequentially. We recommend instructing the agents to give short replies if you plan to connect to Instagram. # Deploy to Shopify Source: https://docs.sketricgen.ai/deploy/shopify Add a SketricGen Brand Agent as an ecommerce AI chatbot for your Shopify store with an inline iframe or floating chat widget. ## **Step 1: Create your SketricGen account** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don't have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. Use this path when you want an ecommerce chatbot on a specific Shopify page, such as Contact, FAQ, product guidance, or demo request pages. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Copy the SketricGen Brand Agent embed code** After configuring your Brand Agent, go to the Brand Agent dashboard and click **Publish > Direct Embed** to copy the iframe embed code. > **Note:** Ensure your Brand Agent is set to **Public** to make it accessible on your Shopify website. Screenshot showing SketricGen Direct Embed iframe code copy interface ## **Step 3: Sign into your Shopify website and embed your Brand Agent** ### 1. Log into your **Shopify** admin dashboard. ### 2. From the left-hand menu, go to **Online Store → Pages**. ### 3. Choose the page you want to edit. (In this example, we're embedding the Brand Agent on the Contact page.) Screenshot showing Shopify admin dashboard with Online Store > Pages navigation ### 4. Paste the embed code into the **HTML/code editor**, then click **Save**. Screenshot showing Shopify page editor with HTML/code editor and Save button ## **Step 4: Test your widget after publishing** Screenshot showing Shopify website with embedded SketricGen chat widget **Success:** Your SketricGen Brand Agent is now integrated into your Shopify site. Visitors can chat with it and receive responses grounded in the knowledge sources you configured. ## **Step 1: Create your SketricGen account** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don't have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. Use this path when you want a floating ecommerce chatbot available across your Shopify storefront. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Copy the SketricGen Brand Agent embed code** After configuring your Brand Agent, go to the Brand Agent dashboard and click **Publish** to copy the embed code. > **Note:** Ensure your Brand Agent is set to **Public** in order to add it to your Shopify website. Screenshot showing SketricGen dashboard with Publish button and embed code copy option ## **Step 3: Sign in to your Shopify account and embed your Brand Agent in your desired theme** ### 1. Click on **Customize theme** Screenshot showing Shopify admin dashboard with Customize theme option ### 2. Add a **Custom Liquid Section** Screenshot showing Shopify theme customizer with Add Custom Liquid Section option ### 3. Paste the code you copied in step 2 in the element block and adjust the position and size of the widget accordingly Screenshot showing Shopify Custom Liquid Section with embed code input and positioning options ## **Step 4: Test your widget after publishing** Screenshot showing Shopify website with embedded SketricGen chat widget **Success:** Your SketricGen Brand Agent is now available as a floating chat bubble on your Shopify site. # Deploy to Slack Source: https://docs.sketricgen.ai/deploy/slack Connect your SketricGen Brand Agent to Slack. ## **Step 1: Sign Into SketricGen and Set Up Your Brand Agent** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don’t have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Connect your Slack Workspace** After configuring your Brand Agent, click the **Connect** button on the Slack integration panel in the **Publish** sidebar. Screenshot showing SketricGen Slack integration panel with Connect button in the Publish sidebar ## **Step 3: Grant Permissions** ### 1. Grant SketricGen the required permissions to access your Slack account and workspace. Screenshot showing Slack permission grant interface for SketricGen workspace access ### 2. After success, you will get a message: **“SketricGen has been successfully added to your workspace.”** ## **Step 4: Check Integration Details** ### 1. If everything is smooth, the **Connect** button in the Slack panel of the Publish sidebar will change to **Manage**. Screenshot showing SketricGen Slack integration panel with Manage button after successful connection ### 2. Click on the **Manage** button to view your integration details. Screenshot showing SketricGen Slack integration details and configuration options ## Step 5: Open Slack and Message Your Bot ### 1. Open any channel in your Slack workspace and type **@sketricgen** followed by a question related to your Brand Agent’s purpose. Slack will then prompt you to either invite the bot to the channel or take no action. Screenshot showing Slack channel with @sketricgen bot mention and invitation prompt ### 2. Click on **Add Them**. The Brand Agent will then be available in the channel to answer any questions you might have. ## Step 6: Start Chatting You're all set! The SketricGen Brand Agent is now connected and ready to enhance collaboration within your Slack workspace. Whenever you or your teammates have a question, simply mention `@sketricgen` along with your query, and the bot will provide an answer. Screenshot showing Slack channel with SketricGen bot responding to user questions ### Success: Slack is connected with your Brand Agent Your team can now ask questions in any Slack channel or DM where the bot is active. The Brand Agent will answer based on the knowledge and instructions you configured in SketricGen. # Deploy to Squarespace Source: https://docs.sketricgen.ai/deploy/squarespace Embed a SketricGen Brand Agent as an AI chatbot for your Squarespace website with a floating chat widget or inline iframe. ## **Step 1: Create your SketricGen account** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don’t have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. Use this path when you want a Squarespace website chatbot for lead capture, support intake, product guidance, or visitor questions. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Copy the SketricGen Brand Agent Embed Code** After configuring your Brand Agent, go to the dashboard and click **Publish** to copy the embed code. > **Note:** Make sure your Brand Agent is set to **Public** to embed it on your Squarespace site. Screenshot showing SketricGen dashboard with Publish button and embed code copy option ## **Step 3: Sign in to your Squarespace account and embed your Brand Agent** ### 1. Sign in to your Squarespace account and open the dashboard ### 2. Select the site you want to embed the Brand Agent on Screenshot showing Squarespace dashboard with site selection interface ### 3. Click **Edit** Screenshot showing Squarespace site editor with Edit button ### 4. Click **Add Block** Screenshot showing Squarespace Add Block interface ### 5. Select **Code** and paste the embed code you copied in Step 2 Screenshot showing Squarespace Code Block configuration interface\\ Screenshot showing Squarespace Code Block editor with embed code input ### 6. Adjust the position of the widget and click **Save** Screenshot showing Squarespace Code Block save and apply options ## **Step 4: Test your widget after publishing** Screenshot showing Squarespace site with embedded SketricGen chat widget **Success:** Your Brand Agent is now live on Squarespace, ready to serve your visitors with custom responses. ## **Step 1: Create your SketricGen account** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don’t have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. Use this path when you want to place a Brand Agent directly inside a Squarespace page section. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Copy the SketricGen Brand Agent Embed Code** From your dashboard, click **Publish > Direct Embed** and copy the iframe code. > **Note:** The Brand Agent must be set to **Public** to work on Squarespace. Screenshot showing SketricGen Direct Embed iframe code copy interface ## **Step 3: Embed your Brand Agent** ### 1. Log in to Squarespace and open dashboard ### 2. Select the site where you want to add the iframe Screenshot showing Squarespace dashboard with site selection interface ### 3. Click **Edit** Screenshot showing Squarespace site editor with Edit button ### 4. Click **Add Block** Screenshot showing Squarespace Add Block interface ### 5. Choose the **Code** block and paste the iframe code Adjust size and placement as needed, then click **Save** Screenshot showing Squarespace Code block with iframe embed and save options ## **Step 4: Test your widget after publishing** Screenshot showing Squarespace site with embedded SketricGen iframe widget **Success:** You've successfully added your SketricGen Brand Agent using an iframe. # Deploy to Telegram Source: https://docs.sketricgen.ai/deploy/telegram Connect your SketricGen Brand Agent to Telegram. ## **Step 1: Sign Into SketricGen and Set Up Your Brand Agent** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don't have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Generate a Bot Token from Telegram** 1. Go to [@BotFather](https://t.me/BotFather) to create a new bot and copy the generated bot token Screenshot showing BotFather interface for creating a Telegram bot and generating a bot token ## **Step 3: Connect the Telegram bot to your SketricGen Brand Agent** 1. Go to the **Publish** section of your Brand Agent and press **Connect** on Telegram Screenshot showing SketricGen Telegram integration panel with Connect button in the Publish sidebar 2. Paste your Bot Token that you copied in Step 2 and press Connect Screenshot showing Telegram bot token input field in SketricGen integration interface 3. Your Connect button will change to Manage and you will see your bot details Screenshot showing SketricGen Telegram integration panel with Manage button after successful connection ## **Success** **Congratulations!** Your Telegram bot is now successfully connected to **SketricGen**. Your Brand Agent will now respond when your Telegram bot receives a query. # Deploy to WhatsApp Source: https://docs.sketricgen.ai/deploy/whatsapp Connect your SketricGen Brand Agent to WhatsApp Business. Integrating WhatsApp with SketricGen enables your custom Brand Agent to engage directly with customers through WhatsApp, offering a streamlined and efficient solution for managing inquiries and automating responses. This guide will take you through the steps required to link your Brand Agent to a WhatsApp number, ensuring smooth and effective customer communication. ## **Before You Begin** The phone number you connect to SketricGen via WhatsApp can **only** be used by the Brand Agent. It will no longer work with the WhatsApp or WhatsApp Business apps. If the number is already in use, you’ll need to delete the associated WhatsApp account first: * Open the **WhatsApp** or **WhatsApp Business** app. * Go to **Settings > Account**. * Tap **Delete my account**. This process may take a few minutes. If you've previously connected the number using Meta's WhatsApp Business tools, make sure to **disable two-step verification**: * Log in to your [**WhatsApp Business Account**](https://business.facebook.com/wa/manage/home/). * Select the phone number you want to integrate. * Go to **Settings > Two-step verification**, then turn it off. ## **Pre-requisite: Sign Into SketricGen and Set Up Your Brand Agent** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don’t have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Go to Agent Playground** Open your Brand Agent and click **Playground** on the agent you want to integrate with WhatsApp. Screenshot showing SketricGen Agent Spaces with Playground button for WhatsApp integration ## **Step 2: Initiate WhatsApp Integration** Click the **Connect** button on the WhatsApp integration panel in the **Publish** sidebar. Screenshot showing WhatsApp integration panel with Connect button in the Publish sidebar ## **Step 3: Set Up with Meta** ### 1. Log in to your Facebook account. Screenshot showing WhatsApp integration setup interface with phone number input field ### 2. Click **Get Started**. Screenshot showing WhatsApp verification code input interface ### 3. Create or select a business portfolio. Screenshot showing WhatsApp integration success confirmation message ### 4. Select an existing WhatsApp business profile or create a new one. Screenshot showing WhatsApp integration configuration settings ### 5. Fill in the information for the WhatsApp business account. Screenshot showing WhatsApp webhook configuration interface ### 6. Add a new number or get a free WhatsApp number by Meta. Screenshot showing WhatsApp Business API configuration panel ### 7. Click **Continue**. Screenshot showing WhatsApp message template configuration interface ### 8. Meta will verify your information. Screenshot showing WhatsApp integration testing interface with message preview ### 9. Click **Finish** to connect your WhatsApp account with SketricGen. Screenshot showing WhatsApp integration status and connection details ## **Step 4: Confirm Connection** If your WhatsApp account is successfully connected, the **Connect** button will turn into a **Manage** button in the **Publish** sidebar of your Brand Agent dashboard. Screenshot showing WhatsApp integration final configuration and deployment options ## **Step 5 (Optional): Manage and Test** You can view integration details and test the setup by clicking the **Manage** button. Screenshot showing WhatsApp integration completion and testing interface ## **Success** **Congratulations!** Your WhatsApp account is now successfully connected to **SketricGen**. You’re all set—your Brand Agent is now ready to handle all your business inquiries automatically and efficiently. Integrating WhatsApp with SketricGen enables your custom Brand Agent to engage directly with customers through WhatsApp, offering a streamlined and efficient solution for managing inquiries and automating responses. This guide will take you through the steps required to link your Brand Agent to a WhatsApp number, ensuring smooth and effective customer communication. ## **Step 1** Navigate to your Brand Agents and open the Brand Agent you want to integrate with WhatsApp. Screenshot showing WhatsApp Business App integration setup interface ## **Step 2** Click the **Connect** button on the WhatsApp integration panel in the **Publish** sidebar. Screenshot showing WhatsApp Business App configuration panel ## **Step 3** ### **1. Login to your Facebook Account** Screenshot showing WhatsApp Business App API settings interface ### **2. Click Get Started** Screenshot showing WhatsApp Business App webhook configuration ### **3. Select Connect your existing WhatsApp Business App** Screenshot showing WhatsApp Business App integration testing interface ### **4. Create or select a business portfolio** Screenshot showing WhatsApp Business App message template setup ### **5. Enter your WA Business phone number** Screenshot showing WhatsApp Business App integration status dashboard ### **6. Verify details** Screenshot showing WhatsApp Business App deployment configuration You will be then asked to scan a QR Code to verify connection. ### **7. After scanning your QR Code, confirm the Permissions** Screenshot showing WhatsApp Business App final setup and testing ### **8. Meta takes a few seconds to verify your information** Screenshot showing WhatsApp Business App integration completion interface ### **9. Click Finish to connect your WhatsApp account with SketricGen** Screenshot showing WhatsApp Business App final configuration and deployment options ## **Step 4** If your **WhatsApp account** is successfully connected with **SketricGen**, the **Connect** button should now turn into a **Manage** button in the **Publish** sidebar in your Brand Agent dashboard. Screenshot showing WhatsApp Business App integration testing and validation ## **Step 5 (Optional)** You can view details of your integration and even test out the integration by clicking the **Manage** button. Screenshot showing WhatsApp Business App integration completion and success confirmation **Success:** Your WhatsApp account is now connected to **SketricGen**. Your Brand Agent can now handle customer inquiries from WhatsApp. # Deploy to Wix Source: https://docs.sketricgen.ai/deploy/wix Connect a SketricGen Brand Agent to Wix with the recommended Wix app, or use a manual floating widget or inline iframe. The **SketricGen Brand Agent** app is the recommended way to add a Brand Agent to Wix. It connects your site without code or an embed snippet and displays the widget across your published site. Manual embed options remain available for custom layouts. ## Before you start You need: * A SketricGen account with at least one Brand Agent. If you do not have one yet, [create a Brand Agent](/brand-agents/create-brand-agent). * Wix permission to install or manage apps and publish the site. A site owner, co-owner, or collaborator with an appropriate custom role can complete the setup. ## Step 1: Make your Brand Agent public and published In SketricGen, open your Brand Agent and select the **Deploy** tab. 1. Set **Visibility** to **Public**. The page confirms that your agent is publicly accessible. 2. Select **Publish Now** if the agent has unpublished changes. SketricGen Brand Agent Deploy tab showing Public visibility and the Publish Now button > **Note:** Wix lists only Brand Agents that are active, published, set to **Public**, and editable by your account. If an agent is missing from the connection list, verify these settings before trying again. ## Step 2: Open the Wix App Market From your Wix site dashboard, go to **Apps → App Market**. Wix site dashboard with Apps expanded and App Market selected ## Step 3: Find the SketricGen Brand Agent app Search for **SketricGen**, then select **SketricGen Brand Agent** from the results. Wix App Market search showing the SketricGen Brand Agent app ## Step 4: Add the app to your site On the app page, select **Add to Site** and accept the permissions requested by Wix. The app is free to install. SketricGen Brand Agent listing in the Wix App Market with the Add to Site button ## Step 5: Start the connection The app appears in your dashboard under **Apps → SketricGen Brand Agent**. Open it, then select **Connect Brand Agent** in **01 Connect your Brand Agent**. A SketricGen window opens. SketricGen Brand Agent setup page in Wix showing the Connect Brand Agent button > **Note:** If the connection window does not open, allow popups for Wix and select **Connect Brand Agent** again. ## Step 6: Sign in to SketricGen Sign in with the account that can edit the Brand Agent you want to connect. Continue with Google or your email address. SketricGen sign-in window opened from Wix with Google and email sign-in options ## Step 7: Choose and connect your Brand Agent Open the **Brand Agent** list and select the agent you want to use. Each entry includes its Teamspace and Project so that you can distinguish agents with similar names. Then select **Connect agent**. SketricGen connection window with a Brand Agent selected from the list After the connection succeeds, the Wix setup page displays **Connected and ready** with the connected agent ID. ## Step 8: Choose the launcher and enable the widget Under **02 Customize the widget**, choose how visitors open the Brand Agent: * **Chat bubble:** A compact launcher in the bottom-right corner. * **Floating pill:** A message field centered along the bottom of the page. Under **03 Publish to your website**, set **Show Brand Agent widget** to **On**. Wix app settings showing the Chat bubble launcher selected and the Brand Agent widget enabled ## Step 9: Publish and test your site Publish your site from the Wix editor, then open the live site. The Brand Agent widget appears with the welcome message configured in SketricGen. Published Wix site with the SketricGen Brand Agent widget open **Success:** Your Brand Agent is live on Wix. ## Customize the widget Appearance, welcome message, suggested questions, knowledge, and behavior are managed in SketricGen. From the Wix app, select **Open Brand Agent Editor**, or open the agent directly at [sketricgen.ai](https://www.sketricgen.ai), and then select **Customize Widget**. Publish changes in SketricGen to send them to the live Wix widget without republishing your Wix site. > **Note:** The launcher style is stored in Wix. After switching between **Chat bubble** and **Floating pill**, publish the Wix site again. ## Change, hide, or disconnect the agent From the app setup page: * **Change agent:** Connect a different Brand Agent to the site. * **Disconnect:** Unlink the agent and remove its widget. The Brand Agent and its content remain in SketricGen. * **Show Brand Agent widget → Off:** Hide the widget without disconnecting the agent. Publish the Wix site after changing any of these settings. ## Troubleshooting **My agent is not in the list.** Confirm that the agent is active, published, set to **Public**, and editable by your account. Return to [Step 1](#step-1-make-your-brand-agent-public-and-published), correct the settings, and reconnect. **I see "Reconnect required."** Select **Change agent**, then choose the agent again to finish secure linking. **The connection window opened blank and closed.** Sign in to [sketricgen.ai](https://www.sketricgen.ai) in the same browser, then retry the connection. **The widget is missing from my live site.** Confirm that **Show Brand Agent widget** is **On**, then publish the Wix site. **I signed in to the wrong SketricGen account.** Select **Sign out** in the connection window, then sign in with the account that can edit the agent. **I cannot install the app or publish the site.** Ask the Wix site owner to install the app or grant you a role that can manage apps and publish the site. ## **Step 1: Create your SketricGen account** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don’t have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. Use this path when you want a Wix website chatbot for lead capture, visitor support, product guidance, or sales questions. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Copy the SketricGen Brand Agent Embed Code** Open your Brand Agent and select **Deploy**. Set **Visibility** to **Public**, publish any pending changes, select **Chat Bubble**, and copy the embed code. SketricGen Brand Agent Deploy tab with Public visibility and Chat Bubble selected ## **Step 3: Embed the Brand Agent in Wix** ### 1. Log in to your Wix dashboard. ### 2. Select the site where you want the Brand Agent. Screenshot showing Wix dashboard with site selection interface ### 3. Go to **Add Elements > Embed & Social > Embed Code**. Screenshot showing Wix Add Elements interface with Embed & Social > Embed Code option ### 4. Paste the copied code and adjust position/size as needed. Screenshot showing Wix Embed Code element with code input and positioning options ### 5. Resize the **Embed Code** element so the full chat widget is visible. Click the embedded widget and adjust the element’s **width/height** (or drag the resize handles) until nothing is clipped and the chat opens at the correct size. GIF showing how to resize the Embed Code element in Wix ### 6. Add an **Anchor** and **Pin** the Brand Agent widget to make it a floating bubble. Screenshot showing Wix Anchor and Pin options for floating chat bubble configuration ## **Step 4: Test the Brand Agent** After publishing, test the Brand Agent on your live site. Screenshot showing Wix site with embedded SketricGen floating chat widget **Success:** Your Brand Agent is live as a floating bubble on Wix. ## **Step 1: Create your SketricGen account** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don’t have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. Use this path when you want to place a Brand Agent directly inside a Wix page section. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Get the Embed Code** Open your Brand Agent and select **Deploy**. Set **Visibility** to **Public**, publish any pending changes, select **iframe/Fullscreen**, and copy the iframe code. SketricGen Brand Agent Deploy tab with Public visibility and iframe Fullscreen available ## **Step 3: Add the iFrame Embed to Wix** ### 1. Go to your Wix site dashboard. ### 2. Select the desired site. Screenshot showing Wix dashboard with site selection interface ### 3. Add an **Embed Code** element: **Add Elements > Embed & Social > Embed Code** Screenshot showing Wix Add Elements interface with Embed & Social > Embed Code option ### 4. Paste your code and adjust dimensions for inline display. Screenshot showing Wix Embed Code element with iframe code and dimension adjustments ## **Step 4: Publish and Verify** Make sure everything is working by testing the Brand Agent on your live site. Screenshot showing Wix site with embedded SketricGen iframe widget **Success:** Your Brand Agent is embedded inline in your Wix site. # Deploy to WordPress Source: https://docs.sketricgen.ai/deploy/wordpress Embed a SketricGen Brand Agent as an AI chatbot for your WordPress website with the SketricGen chat widget plugin. ## **Step 1: Sign Into SketricGen and Set Up Your Brand Agent** To get started with the integration, **[sign in to your SketricGen account](https://www.sketricgen.ai)**. If you don’t have an account yet, you can create one for **[free](https://www.sketricgen.ai)**. Once you're logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. Use this guide when you want to add a WordPress chatbot for visitor questions, lead capture, support intake, or product guidance. If you need assistance with setting up a working SketricGen Brand Agent, refer to the **[step-by-step guide](https://docs.sketricgen.ai/getting-started/quickstart)** for creating and deploying your Brand Agent. ## **Step 2: Install SketricGen Chat Widget on Your WordPress Website** ### 1. Log in to your WordPress admin dashboard You can usually access your dashboard by visiting `yourdomain.com/wp-admin/`. Alternatively, you may be able to log in via your web hosting provider’s control panel. ### 2. Install and Enable the SketricGen Chat Widget * In the left sidebar of your WordPress admin dashboard, click on **Plugins**. Screenshot showing WordPress admin dashboard with Plugins sidebar menu * Click **Add New Plugin** at the top of the next page. Screenshot showing WordPress Add New Plugin interface * In the search bar on the next page, type “**SketricGen**” to search for the **SketricGen** plugin.\ Click **Install Now**, then **Activate**. Screenshot showing WordPress plugin search results for SketricGen with Install Now and Activate buttons ### 3. Get your Agent ID from the SketricGen Brand Agent dashboard Go to your Brand Agent playground and in the **Publish** sidebar copy the Agent ID — it should look like **“asst\_…”** Screenshot showing SketricGen agent playground with Agent ID copy option in Publish sidebar ### 4. Add your Agent ID * In the left sidebar on your admin dashboard, click **SketricGen Chat** under **Settings**. Screenshot showing WordPress admin dashboard with SketricGen Chat settings menu * Enter your **Agent ID**, enable the chat widget, and click **“Save Changes”**. Screenshot showing SketricGen Chat settings interface with Agent ID input field and Save Changes button ## **Step 3: Test your widget on live website** Screenshot showing WordPress website with embedded SketricGen chat widget **Success:** Your WordPress website now has the SketricGen Brand Agent widget integrated. # Deploy to Zapier Source: https://docs.sketricgen.ai/deploy/zapier Integrate your SketricGen Brand Agent with Zapier. ## **Step 1: Sign Into SketricGen and Set Up Your Brand Agent** To get started with the integration, [**sign in to your SketricGen account**](https://www.sketricgen.ai/). If you don’t have an account yet, you can create one for [**free**](https://www.sketricgen.ai/). Once you’re logged in, begin setting up your Brand Agent by uploading your data sources. These can include files, text snippets, websites, or Q\&A pairs, which will serve as the knowledge base for your Brand Agent. If you need assistance with setting up a working SketricGen Brand Agent, refer to the [**step-by-step guide**](https://docs.sketricgen.ai/getting-started/quickstart) for creating and deploying your Brand Agent. ## **Step 2: Copy your SketricGen User ID and API Key** Go to the profile section of your SketricGen account and copy your User ID and API Key which you will paste in Zapier in step 3. Screenshot showing SketricGen profile section with User ID and API Key copy options Screenshot showing SketricGen API Key and User ID details for Zapier integration ## **Step 3: Connect SketricGen with Zapier** ### **1. Create a Zap Workflow** Screenshot showing Zapier dashboard with Create Zap Brand Agent option ### **2. Add SketricGen Integration (We currently support adding our Integration as an Action in a Zapier Workflow)** Screenshot showing Zapier integration marketplace with SketricGen search Screenshot showing Zapier SketricGen integration configuration interface ### **3. Paste the API Key and User ID you copied in Step 2 to connect your SketricGen account with Zapier** Screenshot showing Zapier SketricGen connection with API Key and User ID input fields ## **Step 4: Copy your Agent ID** Go to your Brand Agent playground and in the **Publish** sidebar copy the Agent ID **“asst\_…”** Screenshot showing SketricGen agent playground with Agent ID copy option in Publish sidebar ## **Step 5: Configure your integration** You can now add triggers and actions in your Zap workflow with your SketricGen integration by adding a Prompt Message and your Agent ID. You can also add a Conversation ID so your Brand Agent can maintain context for multi-turn conversations. Your integration will always return a Response Text which you can further use to generate automated replies, generate leads and much more. Screenshot showing Zapier Brand Agent configuration with SketricGen integration, Prompt Message, Agent ID, and Conversation ID fields # Admin API Source: https://docs.sketricgen.ai/dev-guide/admin-api Use the SketricGen Admin API to automate projects, agents, Brand Agents, knowledge bases, conversations, usage, members, and connectors. The SketricGen Admin API is the control plane for managing a Teamspace from your backend, CLI, automation, or custom SDK. It uses the same permissions and plan limits as the SketricGen application. Use the [Runtime API](/dev-guide/public-api-usage) instead when your goal is to execute an existing agent and receive its response. Use the [MCP guide](/dev-guide/mcp-connect) when an AI assistant should manage SketricGen conversationally. ## Base URL ```text theme={null} https://krjavjkt27.execute-api.us-east-1.amazonaws.com/dev/admin/v1 ``` Store it as an environment variable so you can update environments without changing application code: ```bash theme={null} export SKETRICGEN_ADMIN_API_URL="https://krjavjkt27.execute-api.us-east-1.amazonaws.com/dev/admin/v1" ``` ## Authentication Every request requires a Teamspace API key with **Manage resources** access: ```http theme={null} Authorization: Bearer sk_api_YOUR_KEY ``` Create the key from **Teamspace settings → API keys → Create key**. Enable **Manage resources**, choose a Viewer, Editor, or Admin role, and set the Project scope. The plaintext `sk_api_...` value is shown once. If the same key also needs to run agents through the Runtime API or an SDK, enable **Run agents** during the same creation flow. A dual-capability key uses one shared Project scope; agent restrictions apply to its runtime access. When creating the key, choose: * **Project scope:** all accessible Projects, or a selected set of Projects. * **Role:** Viewer, Editor, or Admin. * **Expiration:** 30, 90, 180, or 365 days, or a custom date within the supported maximum. Keys with **Manage resources** access expire and cannot be changed in place. Revoke and recreate a key when its role, scope, capabilities, or expiration needs to change. ## First request Confirm the Teamspace before performing any write: ```bash theme={null} curl "$SKETRICGEN_ADMIN_API_URL/teamspaces" \ -H "Authorization: Bearer $SKETRICGEN_ADMIN_API_KEY" ``` Expected shape: ```json theme={null} { "teamspaces": [ { "teamspace_id": "teamspace_123", "slug": "example-team", "display_name": "Example Team", "subscription_plan": "Builder" } ] } ``` Never log the authorization header or include it in an error report. ## Project and role behavior ### Selected-Project key A selected-Project key can only access its selected Project or Projects. Reads return resources from that scope, and writes targeting another Project return `403 project_scope_mismatch`. ### All-Project key An all-Project key can read across Projects. Write operations that create project-owned resources require a `project_id` in the request body when the key covers more than one Project. ### Roles * **Viewer:** Read Teamspace resources, agents, Brand Agent settings, knowledge bases, conversations, traces, usage, and connector status. * **Editor:** Viewer access plus agent, Brand Agent, knowledge-base, and connector changes. * **Admin:** Editor access plus Teamspace-level project and member management where the endpoint supports it. The effective permission is limited by both the key's role and the key creator's current membership. Removing or demoting the creator can reduce or disable the key immediately. ## Endpoint groups | Area | Available operations | | --------------- | ------------------------------------------------------------------------------------------------------------------------ | | Teamspace | Read the Teamspace associated with the key | | Projects | List, read, create, and rename projects | | Workflow agents | List, create, read, update, delete, and retrieve embed snippets | | Brand Agents | List templates; create asynchronously; read and edit identity, model, instructions, knowledge bases, and widget settings | | Knowledge bases | List, create, read, rename, delete, list data sources, and delete data sources | | Conversations | List conversations, read a conversation, and list its traces | | Members | List members and change an existing member's role with an all-Project key using the Admin role | | Usage | Read plan and credit usage | | Connectors | List connectors and tools, create hosted connection links, check status, attach, and detach | ## Endpoint directory ### Teamspace and projects ```text theme={null} GET /teamspaces GET /projects POST /projects GET /projects/{project_id} PATCH /projects/{project_id} ``` Creating projects requires a key with **Manage resources** access, the **Admin** role, and all-Project (Teamspace-wide) scope. ### Workflow agents ```text theme={null} GET /agents POST /agents GET /agents/{agent_id} PATCH /agents/{agent_id} DELETE /agents/{agent_id} GET /agents/{agent_id}/embed-snippet ``` `POST /agents` expects a complete `workflow_config`. Use Agent Build when a human should visually edit the workflow graph. `POST /agents/{agent_id}/run` currently returns `501 run_not_available`. Run agents through the Runtime API at `POST https://chat-v2.sketricgen.ai/api/v1/run-workflow` with a key that has **Run agents** access. ### Brand Agents ```text theme={null} GET /brand-agent-templates POST /brand-agents GET /jobs/{job_id} GET /brand-agents/{agent_id} PATCH /brand-agents/{agent_id} GET /brand-agents/{agent_id}/widget-config PATCH /brand-agents/{agent_id}/widget-config ``` Brand Agent creation is asynchronous. The initial request returns `202 Accepted` and a `job_id`; poll the job until it reaches `succeeded` or `failed`. ### Knowledge bases ```text theme={null} GET /knowledge-bases POST /knowledge-bases GET /knowledge-bases/{knowledge_base_id} PATCH /knowledge-bases/{knowledge_base_id} DELETE /knowledge-bases/{knowledge_base_id} GET /knowledge-bases/{knowledge_base_id}/data-sources DELETE /knowledge-bases/{knowledge_base_id}/data-sources/{file_name} ``` The Admin API can create an empty knowledge base but does not currently upload files into it. Website-based Brand Agent creation is the supported API flow that crawls content and builds a populated knowledge base. ### Conversations, traces, members, and usage ```text theme={null} GET /conversations GET /conversations/{conversation_id}?agent_id={agent_id} GET /conversations/{conversation_id}/traces?agent_id={agent_id} GET /members PATCH /members/{user_uuid} GET /usage ``` Conversation and trace endpoints are read-only. Member endpoints require a key with **Manage resources** access, the **Admin** role, and all-Project (Teamspace-wide) scope. ### Connectors ```text theme={null} GET /connectors GET /connectors/{app_slug}/tools POST /connectors/{app_slug}/connect-link GET /connectors/{app_slug}/connection POST /agents/{agent_id}/connectors DELETE /agents/{agent_id}/connectors/{app_slug} ``` External-app authorization remains human-in-the-loop. The API returns a short-lived hosted connection URL for the user to open; it never returns the external provider's OAuth token to your integration. ## Example: list agents ```bash theme={null} curl "$SKETRICGEN_ADMIN_API_URL/agents?limit=50" \ -H "Authorization: Bearer $SKETRICGEN_ADMIN_API_KEY" ``` Response: ```json theme={null} { "agents": [ { "agent_id": "agent_123", "name": "Growth Research Agent", "agent_type": "workflow", "agent_status": "active", "project_id": "project_123" } ], "next_token": null } ``` ## Example: create a Brand Agent ```bash theme={null} curl -X POST "$SKETRICGEN_ADMIN_API_URL/brand-agents" \ -H "Authorization: Bearer $SKETRICGEN_ADMIN_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "name": "Example Brand Agent", "seed_url": "https://example.com", "publish_widget": false, "project_id": "project_123" }' ``` The `project_id` field is required for a Teamspace-scoped key and optional for a project-scoped key. Initial response: ```json theme={null} { "agent_id": "skbrand_123", "job_id": "job_123", "status": "queued", "poll_url": "/admin/v1/jobs/job_123", "publish_widget": false } ``` Poll the job: ```bash theme={null} curl "$SKETRICGEN_ADMIN_API_URL/jobs/job_123" \ -H "Authorization: Bearer $SKETRICGEN_ADMIN_API_KEY" ``` Do not treat `queued`, `crawling`, `crawled`, `kb_processing`, or `agent_finalizing` as completion. Only `succeeded` and `failed` are terminal states. ## Example: update a Brand Agent safely Read the current settings first: ```bash theme={null} curl "$SKETRICGEN_ADMIN_API_URL/brand-agents/skbrand_123" \ -H "Authorization: Bearer $SKETRICGEN_ADMIN_API_KEY" ``` Then patch only the intended fields: ```bash theme={null} curl -X PATCH "$SKETRICGEN_ADMIN_API_URL/brand-agents/skbrand_123" \ -H "Authorization: Bearer $SKETRICGEN_ADMIN_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "instructions": "Help website visitors choose the right plan. Ask before collecting personal information." }' ``` Brand Agent patches can change the display name, primary agent name, instructions, model, and assigned knowledge bases. They cannot add, remove, or rewire arbitrary workflow nodes. ## Example: connect an app Connector permissions are a grant to the agent. Use the exact tools published for the connector: 1. `GET /connectors` to choose an app. 2. `GET /connectors/{app_slug}/tools` to inspect grantable actions. 3. `POST /connectors/{app_slug}/connect-link` to create a hosted authorization URL. 4. Give that URL to the intended user. Do not log or persist it. 5. Poll `GET /connectors/{app_slug}/connection` until `connected` is `true`. 6. `POST /agents/{agent_id}/connectors` with the approved `allowed_tools`. Omitting `allowed_tools`, or sending an empty list for some external connectors, can grant every available tool. Enumerate the connector tools and obtain explicit approval for write-capable actions. ## Pagination Most list endpoints use cursor pagination: ```text theme={null} ?limit=50&next_token=OPAQUE_TOKEN ``` * `limit` accepts 1 through 200 and defaults to 50. * Continue while `next_token` is a string. * Stop only when `next_token` is `null`. * Treat the token as opaque; do not decode or modify it. * A short page does not guarantee the listing is complete. The template and connector-tool catalogs use their documented `limit` and `offset` behavior instead of cursor pagination. ## Error format Errors use a stable machine-readable `code`: ```json theme={null} { "error": "Human-readable explanation", "code": "project_scope_mismatch" } ``` Handle the `code` in application logic. The human-readable message may become clearer without changing the code. Common errors: | HTTP | Code | Meaning | | ---- | ------------------------ | ---------------------------------------------------- | | 401 | `missing_token` | No bearer token was supplied | | 401 | `invalid_key` | The key could not be verified | | 403 | `key_expired` | The key is past its expiration | | 403 | `key_revoked` | The key was revoked | | 403 | `project_scope_mismatch` | The request targets a different project | | 403 | `insufficient_role` | The key or its creator lacks the required permission | | 403 | `plan_downgraded` | The active plan no longer includes API keys | | 409 | `workflow_not_ready` | The Brand Agent has not finished provisioning | | 501 | `run_not_available` | Use the Runtime API to execute the agent | ## Operations the Admin API does not expose The Admin API intentionally does not provide: * API-key creation, rotation, or revocation. A signed-in human manages keys in the app. * Billing changes, credit purchases, or subscription cancellation. * Owner transfer or owner-role grants. * New member invitations or member removal. * Direct file upload into a knowledge base. * Provider passwords or raw connector OAuth credentials. ## Related guides * [Connect AI Agents with SketricGen MCP](/dev-guide/mcp-connect) * [Runtime API](/dev-guide/public-api-usage) * [Python SDK](/dev-guide/python-sdk) * [Node.js SDK](/dev-guide/node-sdk) # Connect AI Agents with SketricGen MCP Source: https://docs.sketricgen.ai/dev-guide/mcp-connect Connect Claude, ChatGPT, Codex, and other MCP clients to SketricGen to create, edit, and connect Brand Agents through natural language. Connect an MCP-compatible AI assistant to SketricGen so it can create and manage Brand Agents through conversation. The SketricGen MCP server exposes a focused set of control-plane tools and enforces the same Teamspace, project, role, and plan rules as the SketricGen application. ## Connection details ```text theme={null} Name SketricGen URL https://mcp.sketricgen.ai/mcp Transport MCP over Streamable HTTP Protocol JSON-RPC 2.0 Auth option 1 SketricGen OAuth Auth option 2 Authorization: Bearer sk_api_YOUR_KEY ``` Use **OAuth** for hosted clients that support remote MCP authentication. Use a Teamspace API key with **Manage resources** access for local agents, CLI clients, and custom MCP clients that need a static bearer credential. A key with only **Run agents** access cannot authenticate to the MCP endpoint. MCP management requires OAuth or an API key with **Manage resources** access. ## Which connection path should you use? | AI client | Recommended path | What you provide | | ------------------------------------------------------ | -------------------------------------- | --------------------------------------------------------------------------------------- | | Claude Web or Claude Desktop | Hosted OAuth connector | Only the MCP URL, then SketricGen sign-in | | ChatGPT web developer mode | Remote MCP app with OAuth | MCP URL and OAuth selection | | Codex desktop, CLI, or IDE | OAuth or API key with Manage resources | MCP URL, then login; or an environment variable containing a management-capable API key | | Claude Code, Cursor, Windsurf, or another local client | API key with Manage resources | MCP URL and bearer header | | Custom server-side MCP client | API key with Manage resources | MCP URL and bearer header | Client availability and labels can change. Check the current [ChatGPT developer mode documentation](https://developers.openai.com/api/docs/guides/developer-mode), [Codex MCP documentation](https://learn.chatgpt.com/docs/extend/mcp), or [Claude custom connector documentation](https://support.anthropic.com/en/articles/11175166-about-custom-integrations-using-remote-mcp) when your client shows different menus. ## Option 1: Connect with SketricGen OAuth OAuth is the recommended path for hosted AI clients. The client opens SketricGen sign-in, asks you to choose an accessible project, and requests read and write access. You do not paste an API key, OAuth client ID, or client secret into the AI client. ### Claude Web or Claude Desktop 1. Open **Settings → Connectors** in Claude. 2. For a Team or Enterprise organization, open **Organization connectors** if required by your workspace. 3. Select **Add custom connector**. 4. Name the connector `SketricGen`. 5. Enter `https://mcp.sketricgen.ai/mcp` as the remote MCP server URL. 6. Leave OAuth client ID and client secret empty when Claude offers those fields. 7. Add the connector, then select **Connect**. 8. Sign in to SketricGen, choose the Teamspace project, and approve the requested access. 9. Enable SketricGen from Claude's tools menu in a new conversation. ### ChatGPT web ChatGPT's developer mode is the path for testing a remote MCP server. Availability and workspace approval depend on the current ChatGPT plan and administrator settings. 1. Enable **Developer mode** in ChatGPT settings. 2. Open **Apps** or **Plugins**, then create a developer-mode app. 3. Name the app `SketricGen`. 4. Enter `https://mcp.sketricgen.ai/mcp` as the server URL. 5. Choose **OAuth** authentication. 6. Select **Scan Tools** and complete the SketricGen sign-in and project-consent flow. 7. Create the app after the tool scan succeeds. 8. Start a new chat and select SketricGen from the available apps or developer-mode tools. Do not enter an OAuth client ID or secret unless your client explicitly requires a pre-registered application. SketricGen supports OAuth discovery and Dynamic Client Registration for compatible clients. ### Codex with OAuth Add the following to `~/.codex/config.toml` or a trusted project's `.codex/config.toml`: ```toml theme={null} [mcp_servers.sketricgen] url = "https://mcp.sketricgen.ai/mcp" auth = "oauth" ``` Then start the login flow: ```bash theme={null} codex mcp login sketricgen ``` Complete SketricGen sign-in and project consent in the browser. Use `codex mcp list` or `/mcp` in Codex to confirm the connection. ## Option 2: Connect with an API key Use an API key with **Manage resources** access when the client cannot complete SketricGen OAuth or when a server-side integration needs a non-interactive credential. ### Create the key 1. Open **Teamspace settings → API keys** in SketricGen. 2. Select **Create key**. 3. Enable **Manage resources** and choose the lowest sufficient role: **Viewer**, **Editor**, or **Admin**. 4. Choose all accessible Projects or a selected Project scope. 5. Choose an expiration date. 6. Create the key and copy the `sk_api_...` value immediately. It is shown once. If the same key also needs to run agents through the Runtime API or an SDK, enable **Run agents** during creation and add any agent restrictions. The Project scope is shared across both capabilities. Selected-Project keys implicitly target one selected project when only one is configured. All-Project keys can see multiple projects, so write operations require an explicit project choice. ### Manual MCP client configuration Clients that accept JSON MCP configuration commonly use this shape: ```json theme={null} { "mcpServers": { "sketricgen": { "type": "http", "url": "https://mcp.sketricgen.ai/mcp", "headers": { "Authorization": "Bearer sk_api_YOUR_KEY" } } } } ``` The exact configuration filename and environment-variable syntax differ by client. Store the key outside version control whenever the client supports environment variables. ### Codex with an API key Set the key in your shell or secret store: ```bash theme={null} export SKETRICGEN_API_KEY="sk_api_YOUR_KEY" ``` Reference the environment variable from Codex: ```toml theme={null} [mcp_servers.sketricgen] url = "https://mcp.sketricgen.ai/mcp" bearer_token_env_var = "SKETRICGEN_API_KEY" default_tools_approval_mode = "writes" ``` Restart Codex after changing its configuration. The `writes` approval mode asks before tools that are not marked read-only. ## Verify the connection Start with a read-only prompt: ```text theme={null} Use SketricGen to tell me which Teamspace and project this connection can access. Then list my active agents. Do not change anything. ``` The MCP client should call `whoami` and `list_agents`. Check that the returned Teamspace and project match the scope you intended. Next, test the Brand Agent discovery flow: ```text theme={null} Use SketricGen to show me the available Brand Agent templates and reusable knowledge bases. Do not create an agent yet. ``` ## What your AI can do The SketricGen MCP server currently publishes 15 tools: | Tool | Purpose | | ---------------------------------- | ----------------------------------------------------------------------------------- | | `whoami` | Confirm the connected Teamspace and project scope | | `list_projects` | List projects available to a Teamspace-scoped connection | | `list_agents` | List agents in the authorized scope | | `list_brand_agent_templates` | Discover Brand Agent templates and their required connectors | | `list_knowledge_bases` | Discover reusable, ready knowledge bases | | `create_brand_agent` | Start asynchronous Brand Agent creation from a website | | `get_brand_agent_status` | Follow creation progress and retrieve the finished embed snippet | | `list_brand_connectors` | Discover apps that can be connected to Brand Agents | | `list_connector_tools` | Discover the exact actions that can be granted for one connector | | `create_brand_connector_link` | Create a short-lived link for a person to authorize an app | | `check_brand_connector_connection` | Check whether the person completed authorization | | `attach_brand_connector` | Grant selected connector actions to an agent | | `detach_brand_connector` | Remove a connector from an agent without disconnecting the account | | `update_brand_agent_tool` | Read or change the Brand Agent's identity, instructions, model, and knowledge bases | | `update_brand_agent_config_tool` | Read or change widget appearance, messages, behavior, and visibility | The MCP surface does not currently include a tool for running an agent conversation. Use the [Runtime API](/dev-guide/public-api-usage) or Playground when you want to execute an existing agent. ## Recommended Brand Agent workflow When you ask an AI assistant to create a Brand Agent, it should follow this sequence: 1. Call `whoami` and identify the project silently. 2. Offer available templates before creating anything. 3. Ask for the website URL and offer a reusable knowledge base when one is ready. 4. Call `create_brand_agent` and explain that crawling and knowledge-base creation take several minutes. 5. Poll `get_brand_agent_status` until the status is `succeeded` or `failed`. 6. On success, return the exact embed snippet. 7. Offer required connectors, show their available tools, and ask before granting write-capable actions. 8. Give any connector authorization link to the user to open. The AI should never complete provider sign-in itself. Useful prompt: ```text theme={null} Build a Brand Agent for https://example.com. First show me the available templates. Ask one decision at a time. Before connecting an external app, show me the exact read and write tools it would receive and wait for my approval. ``` ## Security and permission model * **One selected project with OAuth:** OAuth consent binds the connection to one Teamspace project. * **Live membership checks:** Removing or demoting the connected user changes what the OAuth connection or management-capable API key can do. * **Server-enforced roles:** Tool descriptions guide the AI, but the Admin API independently enforces every permission and plan limit. * **Human connector authorization:** The AI receives a short-lived setup link, not your Google, Slack, email, or other provider password. * **Explicit tool grants:** Connector actions are validated against the current connector catalog before they are attached. * **Read before write:** Brand Agent update tools can read current settings before applying a partial update. * **No billing control:** MCP tools cannot change subscriptions, purchase credits, or create additional API keys. An empty connector tool selection can mean unrestricted access for some connectors. Ask the AI to list the connector tools and explicitly confirm any permission that sends, deletes, publishes, spends, or changes external data. ## Troubleshooting ### `401 unauthorized` **Cause:** The client sent no usable OAuth token or API key with Manage resources access. **Fix:** Complete the client's **Connect** or `mcp login` flow. For static authentication, confirm the header is `Authorization: Bearer sk_api_...` and that the key has Manage resources access. ### `403 invalid_origin` **Cause:** A browser-based client sent an `Origin` that the SketricGen MCP server does not currently allow. **Fix:** Use a supported native or hosted client. Contact SketricGen support with the exact client and origin if a browser integration needs to be added. ### `403 insufficient_scope` **Cause:** The OAuth connection approved read access but the requested tool changes data. **Fix:** Reconnect SketricGen and approve write access. ### `403 project_scope_mismatch` **Cause:** A selected-Project API key was asked to act on another project. **Fix:** Use a selected Project in the key's scope, or create a key with all-Project scope when the integration needs multiple Projects. ### `403 key_expired` or `403 key_revoked` **Cause:** The API key with Manage resources access is no longer valid. **Fix:** Create a new key, update the client's secret storage, and remove the old value from local configuration. ### The AI says it changed a Brand Agent, but nothing changed **Cause:** The client may have performed a read or supplied an unsupported field. **Fix:** Ask it to verify that the update tool returned `mode: "write"` and that the intended field appears in `updated` before claiming success. ## For AI agents reading this page * First identify the user's MCP client and whether it supports hosted OAuth. * Prefer OAuth when the client can complete it; do not ask the user for a management-capable API key unnecessarily. * Never ask for the user's SketricGen password or an external app password. * If a static API key is required, instruct the user to create one with Manage resources access in SketricGen and store it in the client's secret or environment-variable mechanism. * Verify the connection with `whoami` before creating or changing anything. * Read existing Brand Agent settings before writing them. * Treat connector setup links as short-lived credentials: show them only to the intended user and do not repeat or store them. * Ask for confirmation before granting tools that write, send, publish, delete, spend, or change data. * Do not claim an asynchronous Brand Agent is ready until `get_brand_agent_status` returns `succeeded`. ## Related guides * [Developer Platform Overview](/dev-guide/overview) * [Admin API](/dev-guide/admin-api) * [Runtime API](/dev-guide/public-api-usage) * [Brand Agent connectors and leads](/brand-agents/connectors-and-leads) # Node.js SDK Source: https://docs.sketricgen.ai/dev-guide/node-sdk SketricGen Node.js SDK — run AI agent workflows, stream responses, upload files, and integrate with TypeScript or JavaScript applications. ## Overview Node.js SDK for the SketricGen Runtime API. Build AI-powered workflows, chat applications, and document analysis tools with TypeScript or JavaScript. ## Requirements * Node.js 18 or later (uses built-in `fetch`, `FormData`, `Blob`, and streams) * ESM: use `"type": "module"` in `package.json` or `.mjs` files * A Teamspace API key with **Run agents** access. Current keys begin with `sk_api_`. Create it from **Teamspace settings → API keys → Create key**. Enable **Run agents**, then choose the Project and agent scope. If the same key also needs to manage resources, enable **Manage resources** as well; a key with only Manage resources access cannot run agents through this SDK. ## Installation ```bash theme={null} npm install sketricgen ``` Or with other package managers: ```bash theme={null} # yarn yarn add sketricgen # pnpm pnpm add sketricgen ``` ## Quick Start ```javascript theme={null} import { SketricGenClient } from 'sketricgen'; // Initialize client (use fromEnv() to read SKETRICGEN_API_KEY from environment) const client = SketricGenClient.fromEnv(); // Run a workflow const response = await client.runWorkflow( 'agent-123', 'Hello, how are you?', ); console.log(response.response); ``` ## Features * **Run Workflows**: Execute chat/workflow requests with AI agents * **Streaming**: Real-time streaming responses using Server-Sent Events * **File Attachments**: Attach images and PDFs to workflows (paths or explicit upload) * **TypeScript**: Full type definitions included * **Error Handling**: Typed error classes (e.g. `SketricGenAPIError`, `SketricGenValidationError`) * **CLI Help**: Run `npx sketricgen --help` or `npx sketricgen --help runWorkflow` for usage *** ## Client Reference ### SketricGenClient The main client for interacting with the SketricGen API. #### Constructor ```javascript theme={null} const client = new SketricGenClient({ apiKey: 'your-api-key', // Required baseUrl: 'https://...', // Optional; default: https://chat-v2.sketricgen.ai uploadInitUrl: 'https://...', // Optional; upload endpoints (not derived from baseUrl) uploadCompleteUrl: 'https://...', timeout: 300, // Optional; request timeout in seconds uploadTimeout: 300, // Optional; upload timeout in seconds maxRetries: 3, // Optional }); ``` #### From Environment Variables ```javascript theme={null} // Set SKETRICGEN_API_KEY environment variable const client = SketricGenClient.fromEnv(); // Optional overrides const client = SketricGenClient.fromEnv({ apiKey: 'override-key' }); ``` | Environment Variable | Default | Description | | --------------------------- | ---------- | ------------------------------ | | `SKETRICGEN_API_KEY` | *required* | Your API key | | `SKETRICGEN_TIMEOUT` | — | Request timeout in seconds | | `SKETRICGEN_UPLOAD_TIMEOUT` | — | Upload timeout for large files | | `SKETRICGEN_MAX_RETRIES` | — | Maximum retry attempts | *** ### runWorkflow() Execute a workflow/chat request. ```javascript theme={null} // Non-streaming: returns Promise const response = await client.runWorkflow( 'agent-123', // agentId (required) 'Hello, how are you?', // userInput (required, max 10,000 chars) { conversationId: 'conv-456', // Optional: resume a conversation contactId: 'contact-789', // Optional: external contact ID filePaths: ['./doc.pdf'], // Optional: local paths; client uploads then attaches assets: ['file-id-1'], // Optional: pre-obtained file IDs from files.upload() stream: false, // Optional: set true for streaming } ); // Streaming: returns AsyncGenerator const stream = client.runWorkflow('agent-123', 'Tell me a story', { stream: true }); for await (const event of stream) { // handle event.event_type, event.data (JSON string) } ``` **Returns**: `Promise` if `stream` is false or omitted; `AsyncGenerator` if `stream: true`. #### Example: Non-Streaming ```javascript theme={null} const response = await client.runWorkflow( 'agent-123', 'What is the weather like today?', { conversationId: 'conv-456' }, ); console.log('Response:', response.response); console.log('Conversation ID:', response.conversation_id); ``` #### Example: Streaming ```javascript theme={null} const stream = client.runWorkflow('agent-123', 'Tell me a story', { stream: true }); for await (const event of stream) { const data = JSON.parse(event.data); if (data.type === 'TEXT_MESSAGE_CONTENT') { process.stdout.write(data.delta || ''); } else if (data.type === 'RUN_FINISHED') { console.log(); } } ``` *** ### files.upload() Upload a file and get a `fileId` for use in `runWorkflow(..., { assets: [fileId] })`. ```javascript theme={null} const result = await client.files.upload({ agentId: 'agent-123', // Required file: './document.pdf', // Required: path (string), Buffer, or Readable stream filename: 'doc.pdf', // Required when file is Buffer or Readable (must include extension) contentType: 'application/pdf', // Optional when file is Buffer/Readable }); console.log(result.fileId); // use in runWorkflow(..., { assets: [result.fileId] }) ``` When `file` is a path string, `filename` and `contentType` are optional (inferred from the path). When `file` is a `Buffer` or Node `Readable` stream, `filename` is required. *** ## Response Models ### ChatResponse Response from a non-streaming workflow request. | Field | Type | Description | | ----------------- | --------- | -------------------------------------- | | `agent_id` | `string` | Workflow/Agent ID | | `user_id` | `string` | User identifier | | `conversation_id` | `string` | Conversation ID for follow-up messages | | `response` | `string` | The assistant's response text | | `owner` | `string` | Owner of the agent | | `error` | `boolean` | Error flag (default: `false`) | ### StreamEvent Individual event from a streaming response. | Field | Type | Description | | ------------ | --------------------- | --------------------------------- | | `event_type` | `string` | Type of the SSE event | | `data` | `string` | JSON string containing event data | | `id` | `string \| undefined` | Optional event ID | Parse `event.data` with `JSON.parse(event.data)` to access the payload. The parsed object may include a `type` field matching `event_type`. *** ## Streaming Events The streaming API uses the [AG-UI Protocol](https://docs.ag-ui.com). Parse the `data` field as JSON to access event details. ### Event Types | Event Type | Description | Key Fields | | ---------------------- | ---------------------------- | -------------------------------- | | `RUN_STARTED` | Workflow execution started | `thread_id`, `run_id` | | `TEXT_MESSAGE_START` | Assistant message started | `message_id`, `role` | | `TEXT_MESSAGE_CONTENT` | Text chunk received | `message_id`, `delta` | | `TEXT_MESSAGE_END` | Assistant message completed | `message_id` | | `TOOL_CALL_START` | Tool/function call started | `tool_call_id`, `tool_call_name` | | `TOOL_CALL_END` | Tool/function call completed | `tool_call_id` | | `RUN_FINISHED` | Workflow completed | `thread_id`, `run_id` | | `RUN_ERROR` | Workflow error occurred | `message` | | `CUSTOM` | Custom event | varies | ### Complete Streaming Example ```javascript theme={null} import { SketricGenClient } from 'sketricgen'; const client = SketricGenClient.fromEnv(); const stream = client.runWorkflow( 'agent-123', 'Search for info and summarize it', { stream: true }, ); for await (const event of stream) { const data = JSON.parse(event.data); const eventType = event.event_type ?? data.type; switch (eventType) { case 'RUN_STARTED': console.error('Started run:', data.run_id); break; case 'TEXT_MESSAGE_CONTENT': if (data.delta) process.stdout.write(data.delta); break; case 'TOOL_CALL_START': console.error('\n[Calling:', data.tool_call_name + ']'); break; case 'TOOL_CALL_END': console.error('[Done]'); break; case 'RUN_FINISHED': console.error('\n\nCompleted!'); break; case 'RUN_ERROR': console.error('\nError:', data.message); break; } } ``` *** ## File Attachments Attach files to workflows for document analysis and image understanding. ### Supported File Types | Type | MIME Types | Max Size | | --------- | ---------------------------------------------------- | -------- | | Images | `image/jpeg`, `image/png`, `image/webp`, `image/gif` | 20 MB | | Documents | `application/pdf` | 20 MB | ### Single File (via filePaths) ```javascript theme={null} const response = await client.runWorkflow( 'agent-123', 'Summarize this document', { filePaths: ['./path/to/document.pdf'] }, ); console.log(response.response); ``` ### Multiple Files ```javascript theme={null} const response = await client.runWorkflow( 'agent-123', 'Compare these two documents', { filePaths: [ './path/to/document1.pdf', './path/to/document2.pdf', ], }, ); console.log(response.response); ``` ### Upload First, Then Run (assets) ```javascript theme={null} const upload1 = await client.files.upload({ agentId: 'agent-123', file: './doc1.pdf' }); const upload2 = await client.files.upload({ agentId: 'agent-123', file: './doc2.pdf' }); const response = await client.runWorkflow( 'agent-123', 'Compare these two documents', { assets: [upload1.fileId, upload2.fileId] }, ); console.log(response.response); ``` ### Streaming with Files ```javascript theme={null} const stream = client.runWorkflow( 'agent-123', 'Analyze this image', { filePaths: ['./path/to/image.png'], stream: true }, ); for await (const event of stream) { const data = JSON.parse(event.data); if (data.type === 'TEXT_MESSAGE_CONTENT') { if (data.delta) process.stdout.write(data.delta); } } ``` # Developer Platform Overview Source: https://docs.sketricgen.ai/dev-guide/overview Choose the SketricGen MCP server, Admin API, Runtime API, or SDK for AI agent management, automation, streaming, and custom applications. SketricGen exposes three developer surfaces. Choose the surface based on whether an AI assistant should manage SketricGen, your software should manage SketricGen resources, or your application should run an agent. A single Teamspace API key can carry runtime access, management access, or both. ## Choose the right integration | Surface | Use it when | Authentication | What it controls | | --------------------------- | --------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------ | | **MCP server** | Claude, ChatGPT, Codex, or another MCP client should build and manage Brand Agents conversationally | SketricGen OAuth or an API key with **Manage resources** access | Brand Agent creation and editing, templates, knowledge bases, widget configuration, and connectors | | **Admin API** | Your backend, CLI, or automation should manage Teamspace resources directly | API key with **Manage resources** access, beginning with `sk_api_` | Projects, agents, Brand Agents, knowledge bases, conversations, traces, members, usage, and connectors | | **Runtime API** | Your application should send a message to an existing agent and receive its response | API key with **Run agents** access, beginning with `sk_api_` | Agent execution through `POST /api/v1/run-workflow` | | **Python and Node.js SDKs** | You want a typed client for the Runtime API | API key with **Run agents** access, beginning with `sk_api_` | Agent execution, streaming, and supported file uploads | ## How the surfaces fit together ```text theme={null} AI assistant or MCP client | | Streamable HTTP + OAuth or sk_api_ key with Manage resources v SketricGen MCP server | | Calls the same authorized control plane v SketricGen Admin API <----- Your backend, CLI, or automation Your product, website, or backend | | API-KEY: sk_api_... with Run agents v SketricGen Runtime API -----> Existing agent run and streamed response ``` The MCP server is an AI-friendly adapter over the Admin API. It does not bypass Teamspace roles, project boundaries, plan limits, or connector permissions. The Runtime API and Admin API remain separate endpoints, but their permissions are capabilities on the same API-key entity. A key can have **Run agents**, **Manage resources**, or both. A key without Run agents cannot execute an agent, and a key without Manage resources cannot create, edit, or delete Teamspace resources. ## Start here * Connect an AI assistant with the [SketricGen MCP guide](/dev-guide/mcp-connect). * Manage resources programmatically with the [Admin API](/dev-guide/admin-api). * Run an existing agent with the [Runtime API](/dev-guide/public-api-usage). * Use the [Python SDK](/dev-guide/python-sdk) or [Node.js SDK](/dev-guide/node-sdk) for a higher-level Runtime API client. ## API key location Create API keys from the active Teamspace: 1. Open **Teamspace settings** in SketricGen. 2. Select **API keys**. 3. Select **Create key**. 4. Enable **Run agents**, **Manage resources**, or both. 5. Choose the narrowest Project scope. If Run agents is enabled, add agent restrictions when needed; if Manage resources is enabled, choose a role and expiry. 6. Copy the plaintext `sk_api_...` key immediately. SketricGen shows it only once. API key creation requires an eligible Teamspace plan and permission to manage API keys. ## Security rule Treat API keys as secrets: * Keep keys on a server or in a secret manager. * Never put a key in a public repository, screenshot, browser bundle, or prompt transcript. * Prefer OAuth for hosted MCP clients because the AI client never needs your management-capable API key. * Use a narrowly scoped key for local agents and automation. * Revoke a key from Teamspace settings when it is no longer needed. # Runtime API Source: https://docs.sketricgen.ai/dev-guide/public-api-usage Use the SketricGen Runtime API to execute existing AI agents, stream responses, continue conversations, attach files, and resume human input. ## Overview The SketricGen Runtime API executes agents that you created in SketricGen. Use it to send a message, continue a conversation, attach previously uploaded files, stream the response, or resume a Human Input pause. The customer-facing Runtime API intentionally exposes one programmatic execution endpoint. Teamspace and agent management belong to the [Admin API](/dev-guide/admin-api), while conversational AI management belongs to the [SketricGen MCP server](/dev-guide/mcp-connect). **Base URL:** `https://chat-v2.sketricgen.ai/api/v1` ## When to use the Runtime API Use the Runtime API when you want to call an existing SketricGen agent from your own product, backend, workflow runner, or custom UI. Common use cases: * Run AI Workforce agents from your own application * Add SketricGen agents to an internal product experience * Build a custom chat UI instead of using the hosted website widget * Trigger an agent from backend jobs or server-side automations * Keep conversation identity aligned with users or contacts in your own system If you only need a copy-paste website chat bubble or iframe, use [Deploy an AI Chatbot to a Custom Website](/deploy/custom-website) instead. ## Authentication Every Runtime API request requires a Teamspace API key with **Run agents** access. Current keys begin with `sk_api_`. Create it from **Teamspace settings → API keys → Create key**. Enable **Run agents**, choose the narrowest Project and agent scope that supports the integration, and copy the plaintext key when it is shown once. You can also enable **Manage resources** if the same credential needs Admin API access. **Required Header:** ```http theme={null} API-KEY: sk_api_YOUR_KEY ``` The key must have **Run agents** access. A key with only **Manage resources** access cannot execute agents through the Runtime API. ## Quick start ```bash theme={null} export SKETRICGEN_RUNTIME_API_KEY="sk_api_YOUR_KEY" curl -X POST "https://chat-v2.sketricgen.ai/api/v1/run-workflow" \ -H "API-KEY: $SKETRICGEN_RUNTIME_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "agent_id": "agent_abc123xyz", "user_input": "Summarize the latest product feedback.", "stream": false }' ``` Keep the Runtime API key on a server or in a secret manager. Do not call the Runtime API directly from public browser code. ## Endpoints ### Run Workflow **POST** `/run-workflow` Execute an AI workflow using a specified agent to generate intelligent responses based on user input. ### Request Headers | Header | Type | Required | Description | | -------------- | ------ | -------- | ---------------------------------------------------------------------- | | `API-KEY` | string | Yes | Teamspace API key with **Run agents** access, beginning with `sk_api_` | | `Content-Type` | string | Yes | Must be `application/json` | ### Request Body | Field | Type | Required | Description | Constraints | | ------------------------------ | ------- | -------- | -------------------------------------------------------------- | ------------------------------------ | | `agent_id` | string | Yes | Unique identifier for the AI agent to use | - | | `user_input` | string | Yes | The user's message or query to process | Max 10,000 characters | | `conversation_id` | string | No | ID of existing conversation to continue | If omitted, creates new conversation | | `contact_id` | string | No | External identifier to associate conversation with a contact | Max 255 characters | | `assets` | array | No | Previously uploaded asset file IDs to include with the message | Array of file ID strings | | `stream` | boolean | No | Enable real-time response streaming | Default: `false` | | `disable_conversation_history` | boolean | No | Skip prior messages when processing this turn | Default: `false` | | `hitl_resume` | object | No | Resume a paused Human Input request | Requires `conversation_id` | ### Example Request ```json theme={null} { "agent_id": "agent_abc123xyz", "user_input": "Help me analyze this quarterly sales data", "conversation_id": "conv_def456ghi", "contact_id": "customer_12345", "stream": false } ``` ## Contact ID and User Grouping Use `contact_id` when your product already knows who the end user or contact is. Conversations sent with the same `agent_id` and `contact_id` are grouped under the same contact in SketricGen. This is useful when you want to: * See all conversations from one customer or app user together * Continue reviewing a user's history across multiple sessions * Connect SketricGen conversations back to your CRM, database, or product user record * Use the same identity model across API, widget, iframe, and other channels Use a stable internal ID, UUID, or hashed value. Do not send emails, phone numbers, names, or other directly identifiable personal data as the contact ID. Good examples: ```text theme={null} customer_12345 user_uuid_550e8400-e29b-41d4-a716-446655440000 customer_hash_abc123def456 ``` Avoid: ```text theme={null} john.doe@email.com 555-123-4567 John Doe ``` ### Identity Fields by Deployment Method The same agent can be deployed through the website widget, iframe, fullscreen link, or Public API. The identity fields change slightly by deployment method: | Deployment method | Agent identifier | Contact identifier | | -------------------- | ---------------------------------------------- | ----------------------------------- | | Runtime API | `agent_id` in the JSON body | `contact_id` in the JSON body | | Chat bubble widget | `data-agent-id` on the script tag | `data-contact-id` on the script tag | | Floating pill widget | `data-agent-id` plus `data-layout="floating"` | `data-contact-id` on the script tag | | Inline iframe | Agent ID in `/iframe-widget/YOUR_AGENT_ID` | `contactId` query parameter | | Fullscreen link | Agent ID in `/fullscreen-widget/YOUR_AGENT_ID` | `contactId` query parameter | API example: ```json theme={null} { "agent_id": "agent_abc123xyz", "user_input": "What's my order status?", "contact_id": "customer_12345", "stream": false } ``` Widget equivalent: ```html theme={null} ``` Iframe equivalent: ```html theme={null} ``` If the visitor is anonymous, omit the contact identifier. The widget will use its browser-stored anonymous contact behavior, while API requests without `contact_id` create conversations without your external user mapping. ## Response Formats The endpoint supports two response modes. ### 1. Standard Response (stream: false) **HTTP Status:** `201 Created` **Response Body:** | Field | Type | Description | | ----------------- | ------- | ----------------------------------------- | | `agent_id` | string | The agent that processed the request | | `user_id` | string | Identifier for the user account | | `conversation_id` | string | ID of the conversation (new or existing) | | `contact_id` | string | External contact identifier (if provided) | | `response` | string | The AI agent's generated response | | `owner` | string | Owner of the agent | | `error` | boolean | Indicates if an error occurred | **Example Response:** ```json theme={null} { "agent_id": "agent_abc123xyz", "user_id": "user_789xyz", "conversation_id": "conv_def456ghi", "contact_id": "customer_12345", "response": "I'll help you analyze your quarterly sales data. Please share the data file or provide the key metrics you'd like me to examine.", "owner": "company_abc", "error": false } ``` ### 2. Streaming Response (stream: true) **HTTP Status:** `200 OK` **Content-Type:** `text/event-stream` When streaming is enabled, the endpoint returns real-time Server-Sent Events (SSE) as the AI processes the request. **Event Types:** | Event Type | Description | | ---------------------- | ----------------------------------------- | | `RUN_STARTED` | Workflow execution has begun | | `TEXT_MESSAGE_START` | AI has started generating a text response | | `TEXT_MESSAGE_CONTENT` | Incremental text content (streaming) | | `TEXT_MESSAGE_END` | AI has finished generating text response | | `TOOL_CALL_START` | AI has started using a tool/function | | `TOOL_CALL_END` | AI has finished using a tool/function | | `RUN_FINISHED` | Workflow execution completed successfully | | `RUN_ERROR` | An error occurred during execution | **Example Streaming Events:** ``` data: {"type": "RUN_STARTED", "timestamp": 1640995200, "thread_id": "conv_def456ghi", "run_id": "run_123abc"} data: {"type": "TEXT_MESSAGE_START", "timestamp": 1640995201, "message_id": "msg_456def"} data: {"type": "TEXT_MESSAGE_CONTENT", "timestamp": 1640995202, "message_id": "msg_456def", "delta": "I'll help"} data: {"type": "TEXT_MESSAGE_CONTENT", "timestamp": 1640995203, "message_id": "msg_456def", "delta": " you analyze"} data: {"type": "TEXT_MESSAGE_END", "timestamp": 1640995210, "message_id": "msg_456def"} data: {"type": "RUN_FINISHED", "timestamp": 1640995211, "thread_id": "conv_def456ghi", "run_id": "run_123abc"} ``` # Python SDK Source: https://docs.sketricgen.ai/dev-guide/python-sdk SketricGen Python SDK — run AI agent workflows, stream responses, and attach files programmatically with async and sync support. ## Overview Python SDK for the SketricGen Runtime API. Build AI-powered workflows, chat applications, and document analysis tools with a Teamspace API key that has **Run agents** access. ## Requirements * Python 3.9+ * Dependencies: `httpx>=0.25.0`, `pydantic>=2.0.0`, `typing-extensions>=4.0.0` * A Teamspace API key with **Run agents** access. Current keys begin with `sk_api_`. Create it from **Teamspace settings → API keys → Create key**. Enable **Run agents**, then choose the Project and agent scope. If the same key also needs to manage resources, enable **Manage resources** as well; a key with only Manage resources access cannot run agents through this SDK. ## Installation ```bash theme={null} pip install sketricgen ``` Or with other package managers: ```bash theme={null} # uv uv add sketricgen # poetry poetry add sketricgen ``` ## Quick Start ```python theme={null} from sketricgen import SketricGenClient # Initialize client client = SketricGenClient(api_key="your-api-key") # Run a workflow response = await client.run_workflow( agent_id="agent-123", user_input="Hello, how are you?", ) print(response.response) ``` ## Features * **Run Workflows**: Execute chat/workflow requests with AI agents * **Streaming**: Real-time streaming responses using Server-Sent Events * **File Attachments**: Attach images and PDFs to workflows * **Async & Sync**: Both async and synchronous API support * **Type Safety**: Full type hints for IDE support * **Error Handling**: Comprehensive custom exception types *** ## Client Reference ### SketricGenClient The main client for interacting with the SketricGen API. #### Constructor ```python theme={null} client = SketricGenClient( api_key: str, # Your SketricGen API key (required) timeout: int = 300, # Request timeout in seconds upload_timeout: int = 300, # Upload timeout for large files max_retries: int = 3, # Maximum retry attempts ) ``` #### From Environment Variables ```python theme={null} # Set SKETRICGEN_API_KEY environment variable client = SketricGenClient.from_env() ``` | Environment Variable | Default | Description | | --------------------------- | ---------- | ------------------------------ | | `SKETRICGEN_API_KEY` | *required* | Your API key | | `SKETRICGEN_TIMEOUT` | `300` | Request timeout in seconds | | `SKETRICGEN_UPLOAD_TIMEOUT` | `300` | Upload timeout for large files | | `SKETRICGEN_MAX_RETRIES` | `3` | Maximum retry attempts | *** ### run\_workflow() Execute a workflow/chat request (async). ```python theme={null} response = await client.run_workflow( agent_id: str, # Agent ID to chat with (required) user_input: str, # User message, max 10000 chars (required) conversation_id: str = None, # Resume a conversation contact_id: str = None, # External contact ID file_paths: list[str] = None, # Files to attach stream: bool = False, # Enable streaming ) ``` **Returns**: `ChatResponse` if `stream=False`, `AsyncIterator[StreamEvent]` if `stream=True` #### Example: Non-Streaming ```python theme={null} response = await client.run_workflow( agent_id="agent-123", user_input="What is the weather like today?", conversation_id="conv-456", # Optional: resume conversation ) print(f"Response: {response.response}") print(f"Conversation ID: {response.conversation_id}") ``` #### Example: Streaming ```python theme={null} import json async for event in await client.run_workflow( agent_id="agent-123", user_input="Tell me a story", stream=True, ): data = json.loads(event.data) if data["type"] == "TEXT_MESSAGE_CONTENT": print(data["delta"], end="", flush=True) elif data["type"] == "RUN_FINISHED": print() ``` *** ### run\_workflow\_sync() Synchronous version of `run_workflow()`. ```python theme={null} response = client.run_workflow_sync( agent_id: str, user_input: str, conversation_id: str = None, contact_id: str = None, file_paths: list[str] = None, stream: bool = False, ) ``` **Returns**: `ChatResponse` if `stream=False`, `Iterator[StreamEvent]` if `stream=True` #### Example ```python theme={null} # Non-streaming response = client.run_workflow_sync( agent_id="agent-123", user_input="Hello!", ) print(response.response) # Streaming for event in client.run_workflow_sync( agent_id="agent-123", user_input="Tell me a story", stream=True, ): data = json.loads(event.data) if data["type"] == "TEXT_MESSAGE_CONTENT": print(data["delta"], end="", flush=True) ``` *** ## Response Models ### ChatResponse Response from a non-streaming workflow request. | Field | Type | Description | | ----------------- | ------ | -------------------------------------- | | `agent_id` | `str` | Workflow/Agent ID | | `user_id` | `str` | User identifier | | `conversation_id` | `str` | Conversation ID for follow-up messages | | `response` | `str` | The assistant's response text | | `owner` | `str` | Owner of the agent | | `error` | `bool` | Error flag (default: `False`) | ### StreamEvent Individual event from a streaming response. | Field | Type | Description | | ------------ | ------------- | --------------------------------- | | `event_type` | `str` | Type of the SSE event | | `data` | `str` | JSON string containing event data | | `id` | `str \| None` | Optional event ID | *** ## Streaming Events The streaming API uses the [AG-UI Protocol](https://docs.ag-ui.com). Parse the `data` field as JSON to access event details. ### Event Types | Event Type | Description | Key Fields | | ---------------------- | ---------------------------- | -------------------------------- | | `RUN_STARTED` | Workflow execution started | `thread_id`, `run_id` | | `TEXT_MESSAGE_START` | Assistant message started | `message_id`, `role` | | `TEXT_MESSAGE_CONTENT` | Text chunk received | `message_id`, `delta` | | `TEXT_MESSAGE_END` | Assistant message completed | `message_id` | | `TOOL_CALL_START` | Tool/function call started | `tool_call_id`, `tool_call_name` | | `TOOL_CALL_END` | Tool/function call completed | `tool_call_id` | | `RUN_FINISHED` | Workflow completed | `thread_id`, `run_id` | | `RUN_ERROR` | Workflow error occurred | `message` | | `CUSTOM` | Custom event | varies | ### Complete Streaming Example ```python theme={null} import json from sketricgen import SketricGenClient client = SketricGenClient(api_key="your-api-key") async for event in await client.run_workflow( agent_id="agent-123", user_input="Search for info and summarize it", stream=True, ): data = json.loads(event.data) event_type = data["type"] if event_type == "RUN_STARTED": print(f"Started run: {data.get('run_id')}") elif event_type == "TEXT_MESSAGE_CONTENT": print(data["delta"], end="", flush=True) elif event_type == "TOOL_CALL_START": print(f"\n[Calling: {data['tool_call_name']}]") elif event_type == "TOOL_CALL_END": print("[Done]") elif event_type == "RUN_FINISHED": print("\n\nCompleted!") elif event_type == "RUN_ERROR": print(f"\nError: {data['message']}") ``` *** ## File Attachments Attach files to workflows for document analysis and image understanding. ### Supported File Types | Type | MIME Types | Max Size | | --------- | ---------------------------------------------------- | -------- | | Images | `image/jpeg`, `image/png`, `image/webp`, `image/gif` | 20 MB | | Documents | `application/pdf` | 20 MB | ### Single File ```python theme={null} response = await client.run_workflow( agent_id="agent-123", user_input="Summarize this document", file_paths=["/path/to/document.pdf"], ) print(response.response) ``` ### Multiple Files ```python theme={null} response = await client.run_workflow( agent_id="agent-123", user_input="Compare these two documents", file_paths=[ "/path/to/document1.pdf", "/path/to/document2.pdf", ], ) print(response.response) ``` ### Streaming with Files ```python theme={null} async for event in await client.run_workflow( agent_id="agent-123", user_input="Analyze this image", file_paths=["/path/to/image.png"], stream=True, ): data = json.loads(event.data) if data["type"] == "TEXT_MESSAGE_CONTENT": print(data["delta"], end="", flush=True) ``` # Structured Output for Developers Source: https://docs.sketricgen.ai/dev-guide/structured-output-tools Define schema-validated JSON tools, detect their public API stream events, and render reliable application UI from SketricGen agent runs. Structured Output is a model-callable tool that emits schema-validated JSON during an agent run. It is useful when your application needs to render a card, save typed project state, trigger routing, or pass a predictable payload to downstream code while the conversation continues across turns. A Structured Output call does not automatically pause the run or end the agent's turn. If your application needs the call to be the final action of the turn, state that in the tool and agent instructions. Use Human Input when the agent must wait for a person's answer. ## Configure the tool In Agent Build: 1. Add **Structured Output** from the advanced tools section. 2. Give it a short title that describes the event, such as `Render recommendations`. 3. Describe exactly when the agent should call it and what the receiving application will do with the payload. 4. Add at least one output field and mark every field your application requires. 5. Connect the tool to the agent that should emit it, save, and test the run. A workflow or template stores the node in this form: ```json theme={null} { "id": "tool-node-recommendations-7d91c2a4", "type": "toolNode", "data": { "name": "emit_render_recommendations_7d91c2a4", "title": "Render recommendations", "toolType": "structured_output", "description": "Research is complete and the frontend needs recommendation cards.", "is_enabled": true, "config": { "version": 1, "output_schema": { "$schema": "http://json-schema.org/draft-07/schema#", "type": "object", "properties": { "templates": { "type": "array", "items": { "type": "object", "properties": { "slug": { "type": "string" }, "reason": { "type": "string" } }, "required": ["slug", "reason"] } } }, "required": ["templates"] } } } } ``` The editor generates `data.name` deterministically from the title and node ID. When building template JSON programmatically, use the current SketricGen template builder instead of inventing this value. ## Use Structured Output for application UI events An application UI event is a use case of Structured Output, not a separate platform tool. Define fields that tell the receiving frontend what to render or do, then detect the normal Structured Output tool-call event in your application. For example, a recommendation workflow can emit component data for cards, while another workflow can emit an event name and the state required to open a form. Keep the schema specific to the agent and application you are building; onboarding-specific forms or events are not global SketricGen tools. Structured Output does not collect the user's response. After the frontend renders an interactive component, send the submitted values as the user's next message on the same conversation. Use Human Input instead if the agent run must pause and wait for that response. ## Schema rules * The root must be a non-empty object schema. * Field names must start with a letter and contain only letters, numbers, and underscores. * Supported field types are string, integer, number, boolean, object, and array. * Arrays can contain supported primitives or objects. * Every name in `required` must exist in `properties` at that level. * Nesting cannot exceed eight levels. * The current contract version is `1`. Keep the payload focused. A small schema is easier for lower-cost models to satisfy and easier for your frontend to evolve safely. ## Detect the event The public streaming API exposes ordinary tool-call events. Match the generated tool name or maintain a node-name mapping in your application. ```ts theme={null} type TemplateRecommendation = { slug: string; reason: string }; function isTemplateRecommendation( value: unknown, ): value is TemplateRecommendation { if (typeof value !== "object" || value === null) return false; const candidate = value as Record; return ( typeof candidate.slug === "string" && typeof candidate.reason === "string" ); } function parseRecommendationResult(result: unknown) { let value = result; if (typeof value === "string") { try { value = JSON.parse(value); } catch { return null; } } if (typeof value !== "object" || value === null) return null; const templates = (value as Record).templates; if (!Array.isArray(templates) || !templates.every(isTemplateRecommendation)) { return null; } return { templates }; } function handleAgentEvent(event: Record) { if ( event.type === "TOOL_CALL_END" && typeof event.tool_call_name === "string" && event.tool_call_name.startsWith("emit_render_recommendations_") ) { const payload = parseRecommendationResult(event.tool_call_result); if (!payload) return; renderTemplateCards(payload.templates); } } ``` `TOOL_CALL_START.tool_call_arguments` contains model-supplied data that has not yet completed runtime validation. Use the start event for progress UI, not application rendering or writes. `TOOL_CALL_END.tool_call_result` contains the validated compact JSON result; depending on the client transport, it may arrive as a JSON string or an object, so parse and runtime-check it before use. Treat the end event as confirmation that this tool call completed, not as proof that the entire agent turn has ended. ## Instruction pattern Tell the agent what must happen before the call and what it must do afterward: ```text theme={null} After you have the user's confirmed goal and a successful template-catalog response, call Render recommendations with two or three catalog slugs. Do not invent slugs. Call it as the final action of the onboarding turn, then stop. ``` Do not ask the model to place raw JSON in prose. Let the tool schema validate the fields and let the frontend own rendering, persistence, and navigation. ## Security and reliability * Validate identifiers such as template slugs against server-trusted data before navigation or writes. * Do not put API keys or private credentials in the output schema, description, or tool arguments. * Make write actions idempotent and require confirmation separately; Structured Output itself only returns data. * Version your frontend component contract inside the payload when UI compatibility matters. * Test missing context, invalid values, duplicate tool calls, interrupted streams, and model retries. The former pre-release shape using `toolType: "custom_tool"` with `config.kind: "structured_output"` is not supported. Use `toolType: "structured_output"` and omit `config.kind`. # Adaptive Onboarding Source: https://docs.sketricgen.ai/getting-started/adaptive-onboarding Use SketricGen adaptive onboarding to analyze your website, understand your goals, and recommend Brand Agents and AI Workforce automations. Adaptive Onboarding is a project-based conversation with the SketricGen onboarding strategist. It learns enough about your business and experience to recommend a useful first agent instead of showing the same setup checklist to everyone. You can provide a public website URL for faster personalization, or describe what you sell, who you serve, and where you operate. Sharing a website is optional. Selecting a role and AI-agent experience before the onboarding strategist asks for business context The walkthrough above stops before any website research begins. The strategist first adjusts the onboarding depth to the user's role and AI-agent experience. ## What Adaptive Onboarding Does Adaptive Onboarding can: * Use your role and AI-agent experience to adjust its explanations. * Review public information from a website you provide. * Identify likely business, website, marketing, or automation bottlenecks. * Research market or growth signals when that evidence can change the recommendation. * Recommend live SketricGen templates that match your goal. * Offer to start Brand Agent setup while the onboarding conversation continues. * Save the conversation and recommendations to the selected Project so you can return later. It does not need to perform a marketing audit for every user. If your need is operational, support-related, data-focused, or unrelated to growth, the strategist can recommend a more relevant workflow instead. ## Complete Adaptive Onboarding 1. Select the role that best matches your work. 2. Select how comfortable you are with AI agents. 3. Start the onboarding conversation. 4. Provide your website URL, or describe your business without a URL. 5. Explain the first problem you want an agent to solve. 6. Review the strategist's observations and recommendations. 7. Choose whether to start a Brand Agent, open a recommended template, or continue to Agent Space. The right side of the onboarding screen shows completed and in-progress activity. The status at the top changes while the strategist is researching or waiting for you. Onboarding strategist asking for an optional website after the user's preferences are saved ## Website and Growth Research When a website is available, the onboarding strategist may inspect: * Public products, services, positioning, audience, and brand language * Website structure and visible conversion paths * Public performance or technical signals when available * Public search, competitor, and market evidence when relevant * Live SketricGen templates that match the identified need The resulting report should distinguish observed facts from third-party estimates and recommendations. Search-volume, traffic, advertising, and competitor data are estimates unless they come from an analytics account you connect separately. Adaptive Onboarding does not automatically gain access to private analytics, CRM data, advertising accounts, or internal documents. Website onboarding starts with the public information you provide. ## Brand Agent Setup in Parallel If a website is available, the strategist can offer to start a website-trained Brand Agent. A Brand Agent can learn the website and brand, answer visitor questions, guide visitors to useful pages, and capture leads. Brand Agent setup does not start without your choice. When you select **Create Brand Agent**, setup opens separately and the onboarding conversation stays available. You can continue discussing your goals while the Brand Agent is prepared. See [Create a Brand Agent](/brand-agents/create-brand-agent) for the complete setup and review process. ## Project Storage and Control The onboarding conversation is scoped to the selected Teamspace and Project. This means: * Switching Projects changes which onboarding state is loaded. * You can leave and later return to the same Project's onboarding conversation. * Skipping onboarding does not delete the saved conversation. * Recommended templates and Brand Agent actions remain choices; they are not executed automatically. ## Troubleshooting ### "We couldn’t prepare your onboarding" SketricGen could not load your Teamspace, Project, or saved onboarding state. Select **Try again**. If the problem continues, select **Skip Onboarding** and open Agent Space, then contact support. ### "Project storage is temporarily unavailable." The strategist may have finished, but the latest conversation state could not be saved. Keep the page open and retry when the storage message clears. Do not restart the onboarding in another Project. ### The website cannot be analyzed Check that the URL is public and accessible without a login. You can continue without it by describing what you sell, your audience, and your main geography. ## Capability Facts * **Product surface:** Adaptive Onboarding * **Primary input:** role, AI-agent experience, optional website, and business goal * **Outputs:** visible recommendations, optional research report, template suggestions, and optional Brand Agent setup * **Can use:** public website analysis, public research, market signals, and the live template catalog * **Cannot access by default:** private analytics, internal files, advertising accounts, or CRM data * **User control:** the user chooses whether to start a Brand Agent, use a template, skip onboarding, or open Agent Space * **Storage scope:** selected Teamspace, Project, and user * **Next step:** [Open Agent Space](/ai-workforce/agent-space) or [create a Brand Agent](/brand-agents/create-brand-agent) # Choose an Agent Type Source: https://docs.sketricgen.ai/getting-started/choose-product Compare public, customer-facing Brand Agents with the private AI Workforce for long-running tasks, multi-agent automation, team collaboration, and files. SketricGen provides two agent types for different environments: * **Brand Agents** are public, customer-facing agents for websites and external channels. * **AI Workforce** agents operate inside a permissioned Teamspace and Project for more powerful internal work. Choose based on who will use the agent and what access it needs—not simply which agent sounds more advanced. ## Quick Decision Choose a **Brand Agent** when customers should talk directly to the agent. Choose **AI Workforce** when you and your team need agents to perform substantial work with files, plans, tools, business systems, and other agents. | Difference | Brand Agents | AI Workforce | | ------------------- | ------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- | | Primary users | Website visitors, prospects, leads, and customers | You and authorized Teamspace members | | Environment | Public, customer-facing deployment | Private, permissioned internal workspace | | Main purpose | Answer questions, guide visitors, capture leads, and represent your brand | Research, create, analyze, automate, and operate across internal workflows | | Work style | Focused customer conversations | Long-running and multi-step work with planning and todos | | Agent structure | One customer-facing agent with controlled tools | A main agent with optional specialist sub-agents and handoffs | | Files and artifacts | Brand knowledge and customer conversation context | Project, agent, and run workspaces with persistent files and generated artifacts | | Collaboration | Teams configure and review the customer experience | Teammates collaborate with the same agents through shared or private conversations | | Tools | Knowledge, lead capture, page navigation, web search, and selected safe actions | Powerful tools, skills, connectors, MCP, API requests, files, code execution, and structured output | | Automation | Responds when customers interact through a deployed channel | Runs manually, through conversations, or automatically on a schedule | ## Choose a Brand Agent for Public Customer Experiences Brand Agents are designed to become the AI layer on top of your website or customer channel. They can learn from your website, build Brand DNA and a Knowledge Base, answer personalized questions, capture leads, and route visitors to the right page. Starting a Brand Agent from a website or manual brand setup Use a Brand Agent for: * A public website chatbot or AI receptionist * Product, pricing, policy, or support questions * Personalized page navigation across a website * Lead qualification and contact capture * Public deployment through website widgets or supported customer channels * Customer conversations that should follow your brand voice Because anyone may interact with a public agent, Brand Agents use a deliberately controlled tool surface. Give them only the knowledge and customer-safe actions they need. ## Choose AI Workforce for Powerful Internal Work AI Workforce is a secure, permissioned agent workspace for you and your team. It is designed for work that may require longer execution, multiple agents, internal tools, reusable files, human decisions, and recurring automation. AI Workforce Agent Space with the Project's agents arranged on a live operational floor If you have used Claude Code or Codex, think of AI Workforce as giving agents a similar cloud workspace for broader business tasks. An agent can work through directories and files, maintain plans and todos, use powerful tools and connectors, create artifacts, and continue substantial tasks in the cloud. SketricGen adds visual Agent Build configuration, multi-agent orchestration, schedules, traces, and Teamspace collaboration around that work. Use AI Workforce for: * Research, content, analysis, sales, support, and operations workflows * Tasks that may run for several minutes and require multiple steps * A main agent that delegates focused work to specialist sub-agents * Project directories, persistent agent files, run files, and generated artifacts * Planning, todos, traces, and human-input checkpoints * Internal connectors, MCP tools, API requests, web search, skills, and code execution * Scheduled reports and recurring automations * Real-time Teamspace collaboration with shared or private conversations AI Workforce conversations in a shared Project are public to authorized Project members by default. You can make a conversation private when only you should be able to view it. ## You Can Use Both Agent Types Many teams use both agent types together: * A Brand Agent acts as the public front door for website visitors and leads. * AI Workforce agents research, create, analyze, and support the internal team behind that customer experience. Both can use approved Knowledge Bases, but their default access and interaction models are different. Brand Agents prioritize safe public engagement. AI Workforce prioritizes powerful, transparent, permissioned work for your team. # SketricGen Core Concepts Source: https://docs.sketricgen.ai/getting-started/core-concepts Understand SketricGen Teamspaces, Projects, Agent Space, Agent Build, Playground, tools, skills, knowledge, runs, files, traces, and human input. SketricGen combines product structure, agent configuration, runtime activity, and human control. These definitions describe how the current platform works. ## Teamspaces and Projects A **Teamspace** contains members, roles, shared plan resources, and one or more Projects. A **Project** scopes agents, conversations, connections, files, onboarding state, and collaboration. Always confirm the selected Project before creating or changing an agent. ## Adaptive Onboarding [Adaptive Onboarding](/getting-started/adaptive-onboarding) is a saved Project conversation that recommends a useful first automation. It can use an optional public website, business goals, public research, and the live template catalog. ## Agent Space [Agent Space](/ai-workforce/agent-space) is the live operational workplace for a Project. It shows agents, active work, schedules, collaborators, and human-input requests. Agent Space does not define the internal agent graph. ## Agent Build [Agent Build](/ai-workforce/agent-build) defines how one agent works. It contains the main agent, optional sub-agents, instructions, models, tools, skills, notes, and connections. ## Playground [Playground](/ai-workforce/playground) is the Run tab inside Agent Build. It manages test conversations, live activity, human input, file attachments, workspace artifacts, and conversation history. ## Main Agents and Sub-Agents The **main agent** receives the user's request, decides how to handle it, uses tools, delegates to sub-agents, and returns the final answer. A **sub-agent** is a specialist called by the main agent. It does not talk directly to the user and cannot have its own sub-agents. ## Instructions and Models **Instructions** define an agent's identity, responsibilities, process, boundaries, tool rules, escalation, and output expectations. The **model** provides the underlying language and reasoning capability. Model selection affects capabilities, speed, cost, context, and supported media. ## Tools, Connectors, and Skills A **tool** performs an action or retrieves information. Examples include File Search, Web Search, API Request, Structured Output, and custom MCP tools. A **connector** is an approved connection to an external app and its actions. Access can be scoped to a Project and may use a managed connection or credentials supplied by your Teamspace. A **skill** is reusable task guidance and supporting resources attached to an agent. Skills improve how an agent performs a type of work; they do not automatically grant access to an external app. ## Knowledge Bases A [Knowledge Base](/shared-platform/knowledge-bases) stores approved website, document, text, or Q\&A content that an agent can retrieve. Knowledge Bases help keep answers grounded in your information. ## Runs and Schedules A **run** is one execution of an agent. It can contain model work, tools, sub-agents, files, messages, and human-input pauses. A **schedule** starts repeatable work at a configured time. Test the agent manually before relying on scheduled execution. ## Files, Workspaces, and Artifacts **Project files** can be available across agents in a Project. An **agent workspace** persists reusable files across conversations for one agent. A conversation or run can also have its own working files. An **artifact** is a file or structured result created during work, such as a report, spreadsheet, document, image, or exported dataset. ## Human Input An agent can pause a run and ask a permitted user for a choice, clarification, confirmation, or approval. Agent Space shows these runs under **Needs attention**. Human input gives people control over uncertain or consequential steps. It is different from a public-channel handoff to a support representative. ## Traces A **trace** records how a run progressed, including agent steps, tool calls, sub-agent calls, errors, latency, and credit usage. Use traces to explain an outcome and improve the Agent Build. ## Brand Agents [Brand Agents](/brand-agents/overview) are customer-facing agents with website and brand context, Knowledge Bases, controlled tools, conversations, leads, and deployment settings. They are separate from the internal AI Workforce experience. # SketricGen Overview Source: https://docs.sketricgen.ai/getting-started/overview Learn how SketricGen turns business context into Brand Agents and AI Workforce agents that use tools, knowledge, schedules, files, and human input. SketricGen is a no-code platform for building AI agents that work with your business context, tools, files, and teammates. You can create customer-facing Brand Agents or an internal AI Workforce. The platform separates **where agents work** from **how agents are configured**: * [Agent Space](/ai-workforce/agent-space) is the live workplace where you see agents, active work, schedules, teammates, and requests needing attention. * [Agent Build](/ai-workforce/agent-build) is the internal configuration surface for instructions, models, sub-agents, tools, skills, and handoffs. * [Playground](/ai-workforce/playground) is the Run experience for testing, conversations, files, artifacts, traces, and human input. ## Start With Your Business Context [Adaptive Onboarding](/getting-started/adaptive-onboarding) can learn your role, AI experience, public website, and current bottleneck. It then recommends a practical first agent or live template. If you provide a website, the onboarding strategist can analyze public business and brand information. When relevant, it can research growth signals and offer to start a Brand Agent while the onboarding conversation continues. You remain in control of which recommendations to use. SketricGen does not automatically connect private accounts, publish content, or start a Brand Agent without your action. ## Brand Agents [Brand Agents](/brand-agents/overview) are customer-facing agents trained on your website, Brand DNA, Knowledge Base, and approved tools. Use a Brand Agent to: * Answer visitor questions from your information * Guide visitors to relevant pages * Capture leads inside a conversation * Support product discovery and customer intake * Deploy a branded widget or fullscreen experience ## AI Workforce [AI Workforce](/ai-workforce/overview) is for internal work across research, marketing, content, sales, support, data, and operations. AI Workforce agents can: * Use built-in tools and connected business apps * Delegate work to specialist sub-agents * Use reusable skills and Knowledge Bases * Work with files and create artifacts * Pause for human input or approval * Run manually or on a schedule * Share conversations and operational state with permitted teammates ## How the Main Surfaces Connect 1. Adaptive Onboarding or a template helps you choose a useful starting point. 2. Agent Space shows the agents available in the selected Project. 3. Agent Build defines how an agent works internally. 4. Playground runs the draft or published agent. 5. Traces, files, and human-input requests show what happened during the run. 6. Runs and schedules repeat useful work after it has been tested. ## Next Steps Get recommendations based on your role, website, and first automation goal. Create, configure, run, and verify a simple AI Workforce agent. Build a website-trained, customer-facing agent. Understand Projects, agents, tools, skills, runs, files, and human input. # Quickstart: Build and Run Your First AI Agent Source: https://docs.sketricgen.ai/getting-started/quickstart Create your first SketricGen AI Workforce agent, configure it in Agent Build, run it in Playground, and verify its work from Agent Space. This quickstart takes you from the selected Project to one completed AI Workforce run. You do not need to connect an external app for the first test. ## Before You Start You need: * A SketricGen account * Access to a Teamspace and Project * Enough Project credits for a short test run * A simple task you can verify yourself If you want personalized recommendations first, complete [Adaptive Onboarding](/getting-started/adaptive-onboarding). You can also skip it and start directly in Agent Space. ## 1. Open Agent Space Select **AI Workforce → Agent Space**. Confirm the Teamspace and Project shown in the switcher. Agent Space showing the selected Project and available agents ## 2. Create an Agent Select **Create new agent**. Choose a suitable template or begin with a simple agent. The screenshots in this quickstart use a Blog Writer and SEO workflow as an example of an internal AI Workforce. It combines a main content-planning agent, research tools, a specialist SEO sub-agent, and workspace files. You can follow the same steps with another internal template. For the first test, use one clear responsibility such as: * Summarize a document * Research a topic and return sources * Turn notes into a structured brief * Answer questions from a Knowledge Base Avoid a multi-agent workflow until one main agent works correctly. ## 3. Configure Agent Build Open **Manage Agent Build** and configure: 1. A descriptive agent name 2. Instructions explaining the result it owns 3. A model appropriate for the task 4. Only the tools or Knowledge Base needed for the test Blog Writer AI Workforce configuration with content strategy instructions, model, tools, and a specialist SEO sub-agent Save the draft before running it. ## 4. Run a Test Select **Run**, then **New Chat**. Enter a realistic prompt with enough detail to judge the result. Playground showing a completed Blog Writer test with conversation history and Project and agent file workspaces Do not test with "hello" alone. A useful test prompt should require the instructions or tool you configured. ## 5. Review the Result Check: * Did the agent follow the requested format? * Did it use the correct tool or Knowledge Base? * Did it delegate only when needed? * Did it create the expected file or artifact? * Did it request human input before an uncertain or consequential step? * Does the trace explain what happened? Change one part of the Agent Build at a time, then repeat the same prompt in a new conversation. ## 6. Return to Agent Space Return to Agent Space and open the agent. You should see its configuration summary, recent activity, internal team, schedules, and any request needing attention. Your first success is complete when: * The agent exists in the correct Project. * A completed conversation appears in Playground history. * The output can be verified against your request. * You can explain the result from the trace or visible activity. ## Next Steps * Add a focused [sub-agent](/user-guides/orchestration-and-handoffs) or a dedicated [Structured Output tool](/dev-guide/structured-output-tools). * Connect [tools, apps, or skills](/user-guides/tools). * Add trusted [Knowledge Base content](/shared-platform/knowledge-bases). * Learn about [runs, schedules, files, and artifacts](/ai-workforce/runs-files-artifacts). # SketricGen Product Capability Reference Source: https://docs.sketricgen.ai/reference/product-capabilities Reference SketricGen product surfaces, AI agent capabilities, inputs, outputs, permissions, controls, limitations, and supported execution paths. This page is the public capability map for SketricGen. It defines what each product surface does, how the surfaces connect, and where user choice, Project permissions, or configuration limit an agent. ## Product Structure * A **Teamspace** contains members, roles, shared plan resources, and Projects. * A **Project** scopes agents, onboarding state, conversations, connections, files, and collaboration. * An **Agent Build** defines the internal behavior of one AI Workforce agent. * **Agent Space** shows live operational state across the agents in a Project. * **Playground** runs and tests an Agent Build through conversations. * A **Brand Agent** is a separate customer-facing agent configured for public conversations and deployment. ## Capability Matrix | Surface | Primary purpose | Inputs | Outputs and visible state | | -------------------- | -------------------------------------------------- | ----------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------- | | Adaptive Onboarding | Recommend a useful first agent or workflow | Role, AI experience, optional public website, business goal | Conversation, optional research report, Brand Agent offer, template recommendations | | Agent Space | Monitor and enter live work across a Project | Selected Project, agent runtime state, presence, schedules | Agent status, active work, teammates, pending human input, operational details | | Agent Build | Configure one internal AI Workforce agent | Instructions, models, main agent, sub-agents, tools, skills, knowledge, connections | Draft and published agent configuration | | Playground | Run and test an agent | Text, files, voice input, human responses | Conversations, agent activity, files, artifacts, human-input requests, traces | | Brand Agents | Serve website visitors and public audiences | Website, Brand DNA, Knowledge Base, approved tools | Customer conversations, navigation, lead capture, widget or channel responses | | Runs / Schedules | Execute and repeat agent work | Agent, instruction, schedule, timezone, optional files | Run state, conversation, files, trace, next scheduled activity | | Knowledge Bases | Ground agents in approved information | Website crawl, supported files, pasted text, Q\&A | Searchable knowledge sources and retrieved context | | Tools and Connectors | Let agents retrieve information or perform actions | Tool configuration, Project connection, action arguments | Tool results, external actions, errors, trace activity | | Skills | Give agents reusable task guidance | Skill instructions and supporting resources | Improved task-specific behavior inside an Agent Build | ## Adaptive Onboarding Adaptive Onboarding may use public website analysis, public search, market or SEO data, and the live template catalog when they can improve a recommendation. It cannot access private analytics, CRM data, advertising accounts, internal files, or connected apps unless those are provided or connected through a separate authorized product flow. Brand Agent setup requires a user choice. Template recommendations are not installed automatically. ## Agent Space Agent Space can show: * All agents in the selected Project * Working, Needs attention, Scheduled, Sleeping, or Runtime unavailable state * Active runs and last activity * Human-input requests * Deployed schedules and active automations * Sub-agent, skill, and tool counts * Teammates viewing or working in conversations Agent Space does not edit an agent's instructions, models, graph, tools, or skills. Those changes belong in Agent Build. ## Agent Build Agent Build can configure: * One main agent * Specialist sub-agents * Agent names, icons, instructions, and models * Tools, app connectors, API actions, MCP tools, and Knowledge Bases * Skills * Agent-to-agent delegation * Input behavior and structured outputs * Draft and published versions * Supported deployment settings A sub-agent cannot have its own sub-agents and does not talk directly to the user. The main agent owns the conversation and final response. ## Playground and Runs Playground can: * Start a new conversation * Continue a saved conversation * Display conversations started from Playground, Agent Space, or schedules * Accept text, supported files, voice input, and file mentions * Show live agent and tool activity * Pause for human input * Preview or download workspace artifacts * Reattach to supported longer-running work * Expose traces for review An agent run uses only the capabilities available to the selected Project and Agent Build. A connected app does not give every agent access to every action. ## Human Input and Approval An agent can pause and request: * A choice from allowed options * Missing information * Confirmation * Approval or rejection * A custom response The request appears in the same conversation and can move the agent into **Needs attention** in Agent Space. The run continues only after an allowed response or cancellation. Human input does not guarantee that every external action is automatically approval-gated. The Agent Build and tool configuration must define where approval is required. ## Files and Workspaces * **Project files** can be read across agents in the selected Project when permitted. * **Agent workspace files** persist across conversations for one agent. * **Conversation or run files** belong to a specific work context. * **Artifacts** are generated reports, documents, spreadsheets, images, data files, or other supported outputs. File scope affects reuse. A conversation file does not automatically become a Project file. ## Tools, Connectors, and Skills * A **tool** is a callable capability. * A **connector** authorizes selected actions in an external app. * A **skill** is task guidance and resources; it does not grant external credentials. * A **Knowledge Base** provides retrievable information; it does not execute an external action. * A **custom MCP server** can expose tools outside the managed catalog, subject to configuration and access controls. SketricGen supports a large connector catalog, but availability, action scope, authentication method, and plan access vary by connector. ## Brand Agents Brand Agents can be configured to: * Learn from public website content and approved Knowledge Bases * Use Brand DNA for identity and response style * Answer visitor questions * Navigate visitors to approved website pages * Capture leads inside chat * Use selected customer-facing connectors * Publish as a widget, fullscreen experience, or supported external channel * Store conversations and leads for review Public agents should receive the minimum safe tool access. A Brand Agent should not be treated as an internal AI Workforce agent with unrestricted business-app access. ## Transparency and Evidence SketricGen exposes different evidence depending on the surface: * Conversation messages show user and agent communication. * Activity views show visible progress and tool work. * Traces show agent steps, tools, sub-agents, errors, latency, and credit usage. * Workspace panels show files available to or generated by a run. * Agent Space shows operational status and pending human work. Model output can still be incorrect. Verify important claims against source material, review generated files, and use human input before consequential external actions. ## Related Documentation * [SketricGen Core Concepts](/getting-started/core-concepts) * [Adaptive Onboarding](/getting-started/adaptive-onboarding) * [Agent Space](/ai-workforce/agent-space) * [Agent Build](/ai-workforce/agent-build) * [Playground](/ai-workforce/playground) * [Brand Agents](/brand-agents/overview) * [Public API](/dev-guide/public-api-usage) # AI Knowledge Bases Source: https://docs.sketricgen.ai/shared-platform/knowledge-bases Create SketricGen AI Knowledge Bases from websites, files, text, Q&A pairs, and editable scraped Markdown for Brand Agents and AI Workforce agents. Knowledge Bases store trusted information that agents can retrieve during a run. They give Brand Agents and AI Workforce agents an approved source of truth instead of relying only on model memory or web search. Both Brand Agents and AI Workforce can use Knowledge Bases, but they use them differently. Knowledge Bases list in SketricGen ## How Knowledge Bases Are Used * **Brand Agents:** usually use Knowledge Bases created from website scraping, support docs, product pages, pricing pages, policies, and FAQs. * **AI Workforce:** use Knowledge Bases for internal docs, SOPs, product information, sales collateral, support policies, and research material. ## Add Sources You can add knowledge through: * Website crawling from a URL * File uploads * Pasted text * Q\&A pairs Supported ingestion types include PDF, DOC, DOCX, TXT, MD, HTML, PPTX, and JSON. ## Website Scraping For Brand Agents, website scraping can create the initial content foundation for a website chatbot. SketricGen reads pages, extracts useful text, and prepares the Knowledge Base for the agent. Review scraped sources before publishing. Add missing files or text when your website does not contain enough detail. Website crawl results inside a Knowledge Base ## Review and Edit Scraped Markdown When a website is scraped, SketricGen stores the crawled page content as editable Markdown. You can open a crawled page, review what was extracted, and clean up the source before the agent uses it. This is useful when a crawl includes: * Repeated navigation or footer text * Outdated content * Missing context from the original page * Copy that should be rewritten before the agent answers from it * Pages that should be removed from the Knowledge Base Better source Markdown usually means better agent answers. Treat the scraped Markdown as the editable source of truth for website knowledge, especially before publishing a Brand Agent to visitors. Review and edit scraped website Markdown ## Use Knowledge in Agents After a Knowledge Base is ready, attach it to the [Brand Agent](/brand-agents/create-brand-agent) or [AI Workforce agent](/ai-workforce/agent-spaces) that needs it. Keep product-critical answers in the Knowledge Base and use Web Search only when freshness matters more than strict brand-owned content. Editable Markdown preview for a crawled Knowledge Base page ## Best Practices * Keep product-critical answers in your own Knowledge Base. * Add pricing, policies, docs, and common objections. * Remove outdated files. * Use Web Search only when freshness is more important than strict brand-owned answers. * Re-test the agent after updating knowledge. # Multi-Agent Orchestration and Handoffs Source: https://docs.sketricgen.ai/user-guides/orchestration-and-handoffs Build multi-agent workflows in SketricGen with a main agent, specialist sub-agents, AI-routed delegation, forced handoffs, and traceable execution. Orchestration defines how agents divide work during an AI Workforce run. In SketricGen, one main agent owns the conversation and can delegate focused tasks to specialist sub-agents. Use orchestration when one agent should coordinate a result that requires different skills, tools, knowledge, or review steps. Keep a single-agent build when one clear set of instructions and tools can complete the work reliably. ## How SketricGen Orchestration Works 1. The user sends a request to the main agent. 2. The main agent decides how to approach the work. 3. The main agent completes work directly or delegates a bounded task to a sub-agent. 4. The sub-agent returns its result to the main agent. 5. The main agent combines the available results and responds to the user. The main agent remains responsible for the conversation and final response. A sub-agent cannot add another sub-agent beneath itself. ## Main Agent The main agent is responsible for: * Understanding the user's request. * Planning the work. * Calling tools and connectors. * Calling sub-agents when specialist help is useful. * Returning the final response to the user. Write main-agent instructions that explain the task boundary, available tools, when to delegate, when to ask questions, and what a successful result looks like. ## Sub-Agents Sub-agents are specialist helpers. The main agent can call them for focused tasks and use their results in the final answer. Good sub-agent jobs include: * Classification * Extraction * Summarization * Research * Redaction * Validation * Short reusable subtasks Give each sub-agent a narrow responsibility, focused instructions, and only the tools and knowledge it needs. Sub-agents cannot have their own sub-agents, so keep the orchestration structure one level below the main agent. ## Add and Connect a Sub-Agent In Agent Build: 1. Add a **Sub-Agent** to the canvas. 2. Give it a clear name that describes its specialty. 3. Configure its instructions, model, tools, skills, and knowledge. 4. Connect it to the main agent. 5. Describe when the main agent should delegate work to it. 6. Save the build and test a prompt that should trigger the sub-agent. 7. Open the trace to confirm which agent ran and what it returned. Avoid creating several overlapping sub-agents. The main agent delegates more reliably when each specialist has a distinct purpose. ## AI-Routed and Forced Handoffs Use the handoff behavior that matches the workflow: * **AI-routed handoff:** The agent decides whether to delegate based on the request, its instructions, and the handoff description. Use this for flexible workflows where different requests need different specialists. * **Forced handoff:** The workflow always continues to the connected agent. Use this for a required sequence, such as research followed by validation or drafting followed by compliance review. Keep handoff descriptions specific. Explain what the receiving sub-agent handles and what result it must return. For a forced handoff, use one clear sequential connection so the path stays deterministic. ## Structured Output Is a Dedicated Tool Agent Build no longer has an agent-level **Output Schema** setting. If an agent must emit predictable JSON, add the dedicated **Structured Output** tool and connect it to the main agent or sub-agent that should call it. Use the Structured Output tool for results such as: * Lead qualification summaries * Research briefs with fixed fields * Extraction from uploaded files * QA or validation reports * Connector-ready payloads * Data rendered by a frontend or consumed by another system Do not use Structured Output for every normal chat response. Keep the agent conversational when a person expects a readable answer. To configure its fields, call conditions, runtime event handling, and schema rules, see [Structured Output for Developers](/dev-guide/structured-output-tools). ## Planning in AI Workforce AI Workforce runs can keep a todo/planning state. This helps longer-running agents break a task into steps and continue work across tools, files, and artifacts. The plan is not a replacement for review. Check final outputs, workspace files, and [traces](/user-guides/traces-and-optimizations) before using results. ## Traces Traces show which tools and sub-agents were used, what failed, how long steps took, and what artifacts were created. Use traces when a run takes the wrong path, misses a handoff, or a sub-agent returns a weak result. ## Orchestration Guidelines * Start with one main agent and add a sub-agent only when the responsibility is meaningfully different. * Give every sub-agent a distinct job and a small tool set. * Tell the main agent when it should delegate and what it must do with the result. * Use AI-routed handoffs for flexible decisions and forced handoffs for mandatory sequences. * Keep user-facing conversation ownership with the main agent. * Test direct requests, requests that need delegation, and requests that should not trigger a sub-agent. * Review traces before expanding the workflow. ## Related * [Agent Build](/ai-workforce/agent-build) * [Runs, Schedules, Files, and Artifacts](/ai-workforce/runs-files-artifacts) * [AI Workforce Tools, Connectors, and Skills](/user-guides/tools) * [Structured Output for Developers](/dev-guide/structured-output-tools) * [Traces and Optimization](/user-guides/traces-and-optimizations) # AI Workforce Tools, Connectors, Skills, and MCP Source: https://docs.sketricgen.ai/user-guides/tools Connect AI Workforce agents to knowledge bases, 2,000+ connectors, MCP servers, APIs, web search, Code Interpreter, image generation, and marketplace skills. Tools and skills turn AI Workforce agents from chat interfaces into AI automation tools that can do real work. They let agents retrieve trusted information, act in connected apps, follow optimized task playbooks, generate artifacts, and complete longer internal work. Brand Agents use a leaner, customer-facing tool surface. For public agents, see [Brand Agent Connectors and Lead Capture](/brand-agents/connectors-and-leads). AI Workforce connector browser with app choices ## AI Workforce Tool Types AI Workforce agents can use: * [**Knowledge Bases**](#knowledge-bases): trusted data from websites, files, text, and Q\&A pairs. * [**SketricGen connectors**](#sketricgen-connectors): 2,000+ app actions for internal workflows. * [**Custom MCP tools**](#custom-mcp-tools): company-specific capabilities exposed through MCP servers. * [**API Request**](#api-request): direct HTTP calls for controlled integrations. * [**Structured Output**](#structured-output): validated JSON for application UI events, typed state, and downstream logic. * [**Web Search**](#web-search): current public information. * [**Code Interpreter**](#code-interpreter): sandboxed Python for calculations, file processing, and generated artifacts. * [**Image Generation**](#image-generation): visual asset generation when the tool is enabled. * [**Marketplace skills**](#skills-marketplace): reusable task packages optimized by other builders. Use the section links above to jump to the tool type you are configuring. ## Knowledge Bases [Knowledge Bases](/shared-platform/knowledge-bases) let agents retrieve trusted source material from websites, uploaded files, pasted text, and Q\&A pairs. In Agent Build, the **File Search** tool connects one or more approved Knowledge Bases to the agent that needs them. The agent can then retrieve relevant passages during a run instead of relying only on its model knowledge. Use Knowledge Bases when the agent needs approved company, product, policy, support, sales, or website context. For AI Workforce, this usually means internal reference data the run should cite or use while preparing an artifact. For Brand Agents, Knowledge Bases power customer-facing answers from approved website and brand content. Knowledge Bases are retrieval tools, not broad app connectors. If the agent needs to take action in another system, pair the Knowledge Base with a connector, API Request, or custom MCP tool. ## SketricGen Connectors SketricGen connectors give AI Workforce agents access to 2,000+ applications. They are useful for internal workflows that need to read or update CRMs, calendars, email, documents, spreadsheets, support systems, project management tools, databases, commerce tools, marketing platforms, and team communication apps. When you add a connector, choose the app first, connect the account, then select only the actions this agent should be able to call. For example, a meeting-prep agent might receive calendar actions to list events and retrieve event details, while a sales operations agent might receive CRM actions to update a record or add a note. Connector action picker showing selected and available actions inside a connected app Action selection is the permission boundary that matters day to day. Review the selected action names, descriptions, configurable properties, and connected account before running sensitive workflows. For practical workflows, combine connectors with skills. A meeting-prep agent can use a calendar connector to fetch meeting details, a research skill to structure investigation, a document skill to produce the brief, and the run workspace to save the artifact. See [Runs, Schedules, Files, and Artifacts](/ai-workforce/runs-files-artifacts) for how outputs are stored. ## Custom MCP Tools Custom MCP tools let you connect a hosted MCP server directly to an AI Workforce agent. Use this when a service, vendor, or internal platform already exposes an MCP endpoint and you want to make selected MCP tools available to an agent through Agent Build. This is different from a single API Request. A hosted MCP can expose a group of tools, and SketricGen lets you discover those tools, choose the ones the agent can use, and attach the allowlisted set to the right agent. The example below uses a public weather MCP endpoint, fetches the available tools, and allowlists only the weather lookup tool for the agent. Custom MCP setup with hosted MCP URL, connection type, headers, and tool discovery To configure a Custom MCP tool: 1. Add a **Custom MCP** tool node. 2. Enter the hosted MCP server URL. 3. Choose the connection type, usually **Streamable HTTP** unless the hosted MCP requires SSE over HTTP. 4. Add headers for authentication, such as an authorization token or API key header. 5. Click **Filter Tools** to fetch the tools exposed by that MCP server. 6. Select only the tools this agent should be allowed to call. 7. Save the tool and connect it to the main agent or the specialist sub agent that needs it. The allowlist matters. If the hosted MCP exposes ten tools but this workflow only needs two, select those two. That keeps the agent easier to control, reduces accidental tool calls, and makes traces easier to review. Use Custom MCP when you have a reusable hosted tool surface, an internal service that already speaks MCP, or a third-party product that publishes an MCP endpoint. Use [API Request](#api-request) when you only need to wrap one REST endpoint. For delegation and routing patterns, see [Multi-Agent Orchestration and Handoffs](/user-guides/orchestration-and-handoffs). ## API Request API Request turns a cURL-style REST endpoint into an API automation tool the agent can call. Use it when you have one endpoint, webhook, lookup, or action that should become available to the agent without building a full connector or MCP server. The configuration defines the request the agent is allowed to make: HTTP method, URL, headers, query parameters, request body fields, and tool instructions. Those fields become the tool schema the agent follows when it calls the endpoint. The example below wraps a public weather forecast endpoint as a `GET` tool, then prompts the builder to test the endpoint before saving. API Request setup with Import through AI, HTTP method, URL, headers, and query parameter controls You can configure API Request in two ways: * **Manual setup:** choose the method, paste the endpoint URL, add headers, define query parameters or body fields, then test the API before saving. * **Import through AI:** paste an API documentation URL, OpenAPI spec, or similar reference and describe the endpoint you want. SketricGen maps the endpoint, headers, parameters, and body fields into a usable tool configuration. In the import flow, paste the API documentation or OpenAPI spec and describe the specific tool you want SketricGen to generate. Import API Configuration dialog for generating an API Request tool from documentation Use the description field to tell the agent when to call the API and what a successful response means. For example, a lead enrichment endpoint should only be called after the agent has a valid email or domain; a support ticket endpoint should only be called after the user confirms the issue should be escalated. API Request works best for controlled endpoints such as webhooks, internal APIs, simple read-only lookups, enrichment calls, REST API automation, or one-off write actions. If the agent needs many actions from the same service, use a connector or Custom MCP instead. ## Structured Output Structured Output lets an agent emit schema-validated JSON when a workflow or application needs predictable data. Add a title, explain exactly when the agent should call it, and define the fields the receiving application expects. Application UI events are one use of the same Structured Output tool, not a separate tool type. For example, a custom frontend can watch for a Structured Output call and use its validated fields to render recommendation cards, open a form, update typed project state, or navigate to the next application step. You can also use Structured Output for classifications, extraction results, connector-ready payloads, and any downstream process that needs predictable fields. The call appears in the run stream so your application can read the arguments and validated result. Structured Output does not automatically pause the run or end the agent's turn. If the application needs the event to be the final action for that turn, state that explicitly in the agent and tool instructions. Use Human Input instead when the agent must pause and wait for a person's answer before continuing. For the workflow JSON contract and stream-handling example, see [Structured Output for Developers](/dev-guide/structured-output-tools). ## Web Search Web Search lets an agent look up current public information when a Knowledge Base is not enough. Use Web Search for market research, recent news, public company information, competitive checks, and time-sensitive facts. For workflows that require approved internal facts, pair Web Search with a [Knowledge Base](#knowledge-bases) and tell the agent which source should win when sources conflict. Web Search tool configuration ## Code Interpreter Code Interpreter gives AI Workforce agents a sandboxed Python environment for calculations, data analysis, file transformation, and artifact generation. Use it when the agent needs to inspect structured data, perform repeatable computation, or create a file rather than only explain an answer in chat. Treat generated files and calculations as run outputs that still require review. Give the agent clear input files, expected output format, and validation criteria, then inspect the resulting artifact and trace before using it downstream. ## Image Generation Image Generation lets agents create visual assets when the capability is enabled for your workspace. Use it for workflows that produce social creatives, thumbnails, concept visuals, or supporting images. Treat image-heavy workflows as something to test end to end: the model, tool availability, artifact handling, and any upload connector all affect reliability. ## Skills Marketplace AI Workforce skill marketplace with search, verified skills, install counts, sources, categories, and install buttons Marketplace skills are reusable task packages you can install onto an agent. They make the skills marketplace a practical AI tools marketplace for proven processes such as PDF work, financial analysis, dashboard creation, product requirements, design review, research, writing, or planning. Use skills when you want more than raw app access. A connector lets an agent act in an app. A skill gives the agent a curated way to perform a task well. When browsing skills, review: * The skill name and description. * The source or namespace. * Verification and reputation signals such as installs. * Category tags that match the work you want the agent to do. * Whether the skill should be installed on the main agent or a specialist sub agent. Click **Install** to attach the skill to the selected agent. After installation, the agent can use that skill during runs alongside its instructions, model, files, tools, and connected apps. Installed AI Workforce skills with workflow, version, status, edit, and uninstall controls Installed skills can be reviewed from the Skills page. This helps teams see which workflows use which skills, check status, edit skill files when needed, and remove skills that should no longer be available. ## Tool Assignment Attach tools only where they are needed. The main agent should have the tools required to plan and coordinate the run. Sub agents should receive the smaller tool set required for their specialist task. For example, a research sub-agent may need Web Search and a Knowledge Base, while a reporting sub-agent may need only run files and the dedicated [Structured Output tool](/dev-guide/structured-output-tools). For debugging tool calls, action errors, latency, and credit usage, see [Traces and Optimization](/user-guides/traces-and-optimizations). # Traces and Optimization Source: https://docs.sketricgen.ai/user-guides/traces-and-optimizations Use SketricGen traces to inspect AI Workforce runs, sub-agent calls, tool calls, latency, credit usage, generated files, artifacts, and errors. Traces show what happened during an agent run. They are useful for both Brand Agents and AI Workforce, but they are especially important for longer internal runs. AI Workforce run trace view showing user messages, skills, agent steps, tool calls, and output events ## What Traces Show Traces can show: * Which agent handled the message * Sub-agent calls * Tool calls and tool results * Errors or failed calls * Latency and duration * Credit usage * Structured data * Generated artifacts or file-related events Detailed trace view showing user input, tool result JSON, timing, and step-level run details ## Brand Agent Review For Brand Agents, use traces and conversations to understand why the agent answered a visitor a certain way, whether the Knowledge Base was enough, and whether lead capture or navigation tools fired correctly. ## AI Workforce Review For AI Workforce, use traces alongside run files and artifacts. Longer runs may involve planning, multiple tool calls, generated files, connector actions, marketplace skills, and scheduled automation. ## Optimization Loop Use this loop: 1. Run a realistic test. 2. Open the trace. 3. Identify unclear instructions, slow tools, missing knowledge, or bad routing. 4. Update instructions, tools, Knowledge Bases, or model selection. 5. Re-run the same test. Small changes usually beat large rewrites. For related setup, see [AI Workforce Tools, Connectors, and Skills](/user-guides/tools), [Runs, Schedules, Files, and Artifacts](/ai-workforce/runs-files-artifacts), and [Multi-Agent Orchestration and Handoffs](/user-guides/orchestration-and-handoffs).