429 Research queue full

The organization already has the maximum number of queued asynchronous research jobs (25). A queued job runs within a few minutes; the cap stops one burst from holding the shared research queue.

Safe to retry after a backoff

What causes it

• Many POST .../research?mode=async calls in a short time, faster than the queue drains.

How to fix it

• Poll GET /research-jobs/{id} for the jobs you already queued, then retry after the `Retry-After` interval.
• Call without mode=async to research one record synchronously.

Example response

HTTP/1.1 429
Content-Type: application/problem+json

{
  "type": "https://sealdeal.ai/errors/research-queue-full",
  "title": "Research queue full",
  "status": 429,
  "detail": "..."
}

`type` is stable and safe to branch on in code. `detail` is specific to your request and is the field to read for the particulars; its wording may change.