AI tools that support the Model Context Protocol can now work on your Cosmoner project directly. Connect one to:
https://api.cosmoner.com/v1/mcp
with a project API key, and your assistant can answer "why is my app failing?" by reading its build and runtime logs itself, redeploy it once the fix is pushed, and set the variables it needs. In Claude Code:
claude mcp add --transport http cosmoner https://api.cosmoner.com/v1/mcp \
--header "Authorization: Bearer $COSMONER_API_KEY"
What it can do
- Look around: list projects, apps, servers and variables, and see which secrets exist — never their values.
- Investigate: read an app's configuration, status, deployments, and the latest build or runtime log lines.
- Act: rebuild and redeploy an app, and create or update variables.
It cannot delete anything or create anything you would be billed for.
You decide what it may do
Every action runs with the API key you give it, so it can do exactly what that
key is allowed to do. Create a key just for your assistant with only the scopes
it needs — apps:read to let it read logs, apps:write as well to let it
redeploy — and revoke it from the control panel whenever you like.
See Connect AI Agents for setup in other tools and which scope each action needs.