Docs for developers and agents
HeyCMO reads your public site into a knowledge base, plans your content, and drafts pages and posts in which every claim links to its source. A person approves each piece with a passkey (Face ID, fingerprint or your phone) before it goes out. Agents and scripts can do the rest through three doors: the MCP server, the heycmo CLI and the REST API.
-
Connect over MCP
Add the HeyCMO MCP server to Claude, Cursor or another MCP client. The 14 tools, 2 prompts and 3 resources, and how sign-in picks a workspace.
-
CLI quick start
Run the free audit, sign in, start a job, follow it and ask to publish, from a terminal or a CI job. Output for scripts and the exit codes.
-
Error codes
Every problem code the API can answer, what it means, if a retry helps, and who can clear it.
The REST API
The API is at https://api.heycmo.ai/v1. Its OpenAPI 3.1 description is at heycmo.ai/openapi.json. The MCP server and the CLI call the same API with the same rules, so they refuse the same requests.
- Send a bearer token: a CLI login or an API key that a workspace owner made. Use an MCP sign-in on the MCP server.
- Send an
Idempotency-Keyheader with each write, so the same request sent again does the work once. - Errors are
application/problem+json(RFC 9457). Each answer has acodeand adocs_urlthat opens its section on the error codes page.
What only a person can do
No agent, API key, MCP sign-in or CLI login can do these things. A person does them on app.heycmo.ai, most often with a passkey:
- Approve or reject a piece, or a week of pieces.
- Stop a scheduled post.
- Raise a level or resume a paused workspace.
- Connect a platform, such as GitHub, Webflow or X.
- State a fact, mark a fact wrong, add a site to read, or change the tracked questions.
- Change the plan, invite members, and make or revoke API keys and client access.
When a request needs one of these steps, the answer gives a link to app.heycmo.ai. Give that link to a person. An agent can still lower a level or pause everything, because that is the safe direction.