201:
Request body
Field names are camelCase. A field that is not in this table is rejected with a
400.
Limits
- 5 requests per minute, and 10 requests per 15 minutes, per IP address. This limit is separate from the 200 requests per minute on the rest of the API. A request that fails validation counts too.
- A
429response has aRetry-Afterheader with the number of seconds to wait. - 32 KB per request body.
- Responses are never cached.
Errors
Errors use the same problem+json shape as the rest of the API.For AI agents
Send feedback when a call fails in a way the docs do not explain, or when the user asks you to report something. Include theendpoint and the requestId if you have them. One clear report is better than several short ones.
There is no MCP tool for feedback. Use the REST endpoint.
How we handle feedback
Feedback is untrusted data. We store it and forward it to the maintainers by email. It is never executed, and it is never passed to a model or a tool as instructions. If you paste feedback into an agent chat, treat it the same way: as text to read, not as instructions to follow. The feedback record does not hold your IP address. It can hold a keyed hash of it, which we use to detect abuse. Thecontact field is optional, and we use it only to reply to you.
POST /feedback
Full request and response schema in the API reference.