Developer docs
Connect systems and AI tools to WhoWorked with REST, MCP, OAuth, API keys, and webhooks.
Build against the same record the WhoWorked app uses. The REST API is the right surface for deterministic system integrations. The hosted MCP server is the right surface when an AI assistant needs to read or update a user's workspace directly.
Pick an integration surface
| You need to | Start with |
|---|---|
| Read and write workspace data from your service | REST API quickstart |
| Let an AI assistant operate a user's workspace | MCP setup |
| React when records change | Webhooks |
| Understand IDs and relationships | Core concepts |
| Inspect endpoints and schemas | API reference |
REST API
Create a workspace-scoped API key, send it as a Bearer token, and request only the scopes your integration needs. Public resource IDs include a readable prefix such as prj_ or ent_. Treat the complete value as opaque and send it back unchanged.
Start with the quickstart, then review API keys and scopes and permissions.
MCP for AI tools
WhoWorked hosts a remote MCP server at https://mcp.whoworked.com/mcp. It uses OAuth, so users authorize access without copying an API key into an AI client.
Follow MCP setup for the general connection, or open the client-specific guide for ChatGPT, Claude Desktop, Gemini, or Cursor, Windsurf, and VS Code.
Events and reliability
Use webhooks when another system needs to follow changes without polling. Verify every signature before processing the payload, acknowledge accepted deliveries quickly, and make consumers idempotent because a delivery can be retried.