Set up a Robot from scratch in about five minutes. This guide walks through every field so you know what each one does. ## Before you start - Yao Engine is running and Yao Agents is connected - You have at least one AI model configured (Settings → Models) - You know what you want the Robot to do — even a rough description is enough to get started --- ## Step 1 — Open Mission Control and add a Robot Open **Mission Control** from the left sidebar. You'll see the list of existing Robots (empty if this is your first). Click **+ Add Task Agent** in the top-right corner.  --- ## Step 2 — Fill in the Basic tab The **Basic** tab sets the Robot's identity and operating mode.  **Name** — give it a clear, descriptive name. This is what you'll see in the list and in logs. Examples: "Daily Market Digest", "Weekly Report Builder", "Support Inbox Triage". **Workspace** — choose which sandbox environment the Robot runs in. Each workspace has its own files and context. Use separate workspaces if you want this Robot's files isolated from your chat Experts. **Work mode** — two options: - **Autonomous** — runs on a schedule or event trigger, no human required - **On Demand** — you manually kick it off each time; useful for jobs you want to control when they start **Reports to** — who receives the Robot's output. This can be a user account or another Robot. Setting this correctly ensures the result lands in the right place. > 💡 If you're not sure which work mode to pick, start with **On Demand**. You can switch to Autonomous once you've tested the Robot and are happy with the results. --- ## Step 3 — Configure the Identity tab The **Identity** tab defines what the Robot does and what it has access to.  **Role description** — describe the Robot's job in plain language. Be specific about: - What it's responsible for - What inputs it works with (email subjects, URLs, file types, etc.) - What output it should produce (a report, a summary, an email, etc.) Not sure where to start? Click **Generate** — it will draft a role description based on the name you entered. **AI Experts** — select which AI Experts this Robot can use during its pipeline. A Robot analyzing market data should have Market Scout. One that sends email reports should have Postman. **Tools** — select additional tools the Robot can invoke (web search, file reader, API connectors, etc.). > 💡 **Match the resources to the role.** Think of it like hiring: you wouldn't ask someone to do financial analysis without giving them access to the numbers. A clear role description with the right Experts and tools is the difference between a Robot that works reliably and one that struggles. If the Robot needs to send email, give it Postman. If it needs to search the web, add the search tool. --- ## Step 4 — Save and run Click **Save Changes** when you're done. The Robot appears in the Mission Control list. To run it for the first time: - Select the Robot in the list to open its Active view - Click **Assign Task** — a panel slides in on the right where you describe the task - Robot Host reads your description, confirms the goal, and starts the pipeline once you approve Watch the pipeline progress in real time. Each stage updates as it completes. → [Progress & Results](/docs/en-us/mission-control/progress-and-results) --- ## Step 5 — Set up integrations (optional) Open the **Integrations** tab to connect the Robot to messaging platforms or expose it as an API.  **Messaging platforms** — connect to WeChat, Telegram, Discord, DingTalk, or Lark. Paste the platform credentials and users can chat with the Robot directly from that app. **API Access** — each Robot automatically gets an OpenAI-compatible Chat Completions endpoint. Use it to trigger the Robot from any OpenAI SDK, automation tool, or internal system. --- ## Advanced settings (optional) The **Advanced** tab gives you finer control once the basics are working.  - **Additional Recipients** — add email addresses beyond the default "Reports to" recipient - **Concurrency** — set max concurrent runs and queue size for high-volume Robots - **Execution Pipeline** — replace the default agent for any pipeline stage (Host Agent, Goals, Tasks, Delivery) with a custom one - **Developer Options** — push results to a Webhook URL or call a Yao Process on delivery → [Replacing Stage Agents](/docs/en-us/mission-control/execution-pipeline/replacing-stage-agents) --- ## Troubleshooting **The Robot runs but produces poor results** Revisit the Identity tab. The role description is the most important factor — make it more specific. Also check that the AI Experts and tools you've assigned match what the task actually needs. **The Robot doesn't send output anywhere** Check the **Reports to** field in the Basic tab and make sure Postman or the relevant delivery Expert is listed in the Identity tab. **I triggered the Robot but nothing happened** If the work mode is Autonomous, it only runs on the configured schedule. Switch to On Demand or trigger it manually via the dashboard. ## What's next - [Execution Pipeline](/docs/en-us/mission-control/execution-pipeline/pipeline-overview) — understand what happens inside the pipeline when a Robot runs - [Scheduling & Triggers](/docs/en-us/mission-control/scheduling-and-triggers) — set a schedule or configure event-based triggers - [Progress & Results](/docs/en-us/mission-control/progress-and-results) — monitor runs and access deliverables
Creating a Robot
feedback
反馈