Rate limits
When a limit is reached, the server responds with HTTP
429 and a body that says how long to wait:
retry_after is in seconds. Requests authenticated with an API key return {"error": "Rate limit exceeded", "retryAfter": 1200, "remaining": 0} instead, and the per-IP limit returns {"msg": "…", "retryAfterSeconds": 30, …}.
Wait the indicated time before retrying.
Credits
Evermuse MCP access is included with your Evermuse workspace and is metered in Evermuse Credits:- Each successful tool call uses 1 credit.
find_skillsandread_skillsare free.- Calls that fail aren’t charged.
- Sources you add with
add_sourcealso use credits when Evermuse processes them, the same as uploads in the app. Re-submitting a meeting or communication that already exists isn’t processed again.
Size limits
When a page is trimmed to fit, the result reports
next_offset. Call the tool again with offset set to that value to
continue.
Protocol details
HTTP errors
On a
401 with invalid_token, clients should refresh the access token and retry once. If the refresh fails, sign in
again.
Tool errors
Problems with a specific tool call are returned as a normal MCP result withisError: true and a message the agent can
act on, rather than an HTTP error. Common examples:
Removed or renamed tools return a message explaining the change, naming a replacement where one exists. For example,
see_updated_roadmap points to
get_opportunities.