Agent Orchestrator lets you connect Userorbit with AI agent platforms so that autonomous agents can read and write data in your Userorbit workspace. This is useful for automating support workflows, generating announcements from CI/CD events, analyzing feedback at scale, and more. This tutorial walks you through connecting your first AI agent orchestration platform to Userorbit.
What is Agent Orchestrator?
Agent Orchestrator is a Userorbit feature that exposes your workspace data through a structured interface designed for AI agents. Instead of building custom API integrations from scratch, you can connect platforms like LangChain, CrewAI, AutoGen, or any MCP-compatible orchestrator to Userorbit using pre-built connectors.
Agents connected through Agent Orchestrator can:
- Read announcements, feedback, roadmap items, and help center articles.
- Create and publish announcements.
- Respond to feedback submissions.
- Search and analyze user data.
- Trigger workflows based on Userorbit events.
Step 1: Generate an API key
Your AI agent needs an API key to authenticate with Userorbit:
- Go to Settings > API Keys in your workspace.
- Click Create API Key.
- Name it something descriptive like "Agent Orchestrator - Production."
- Select the appropriate permission level. For most agent use cases, Read and Write is sufficient.
- Copy the key and store it securely — you will need it in the next step.
Step 2: Configure your orchestration platform
The exact steps depend on which platform you are using, but the general pattern is the same:
- In your orchestration platform, add a new tool or integration.
- Set the base URL to
https://api.userorbit.com/api/v1. - Configure authentication with your API key in the
Authorization: Bearer YOUR_API_KEYheader. - Add the
x-team-id: YOUR_TEAM_IDheader. You can find your team ID in Settings > General.
MCP-compatible platforms
If your orchestration platform supports the Model Context Protocol (MCP), Userorbit provides an MCP server configuration. In your MCP client config, add Userorbit as a tool server with the API base URL and authentication headers. The MCP interface automatically exposes available actions like reading feedback, creating announcements, and searching articles.
Custom integrations
If your platform does not have a pre-built Userorbit connector, you can use the REST API directly. Define tool functions in your agent that make POST requests to Userorbit API endpoints. Common endpoints include:
articles.list— List help center articles.articles.create— Create a new article.announcements.list— List announcements.announcements.create— Create a new announcement.feedback.list— List feedback submissions.
Step 3: Test the connection
Before deploying your agent to production, verify the integration works:
- Have your agent perform a simple read operation, like listing recent announcements.
- Verify the response contains the expected data.
- Test a write operation in a staging environment, like creating a draft announcement.
- Check the Userorbit dashboard to confirm the draft was created.
Step 4: Set up guardrails
AI agents operating autonomously should have appropriate guardrails:
- Use the most restrictive API key permissions that still allow the agent to function.
- Configure your agent to create drafts rather than publishing directly, so a human can review before content goes live.
- Set rate limits in your orchestration platform to prevent runaway API calls.
- Monitor agent activity in the Userorbit audit log under Settings > Activity.
Example use cases
- Auto-triage feedback — An agent reads new feedback submissions, categorizes them, and assigns them to the right team.
- Release notes from commits — An agent monitors your Git repository and drafts announcements summarizing each release.
- Support article generation — An agent identifies common support questions from feedback and drafts help center articles to address them.