Rate Limits

Minds enforces usage quotas — not per-request rate limits — on resource creation and question volume. Quotas are checked at the moment you try to create a mind, create a datasource, or send a question (via /chat/completions or /responses/). If you're over your current limit, the request returns 429 Too Many Requests.

What gets checked

Name
Mind creation
Description
Every POST /minds/ call runs a LimitsService check before the mind is created. If you've hit your allowed mind count, the request returns 429.
Name
Datasource creation
Description
Every POST /datasources/ call runs the same check against your datasource quota.
Name
Question submission
Description
Calls to /chat/completions and /responses/ check your question quota before routing to the agent. Exceeding it returns 429.

Checking your current limits

Call GET /limits/ with your API key to retrieve the authenticated user's current usage limits:

cURL

curl --request GET \
  --url 'https://mdb.ai/api/limits/' \
  --header 'Authorization: Bearer MINDS_API_KEY'

The response is a MindLimitsConfig object describing your current quotas.

Handling 429s

When you hit a limit:

Read the detail field from the response body — it identifies which quota was exceeded. See Errors.
Decide whether to retry later or request a higher limit (below).
Implement backoff on the calling side rather than retrying immediately.

Requesting a higher limit

If your workload needs more capacity than the default quotas allow, contact us via the form below.