Guides
Idempotency
Protect retried logical operations from duplicate usage charging.
Idempotency-Key
Reuse an Idempotency-Key only when retrying the same logical operation. It protects duplicate usage charging; it does not identify a gateway request, cache a response, or guarantee that an upstream Provider call is skipped.
cURL
1curl https://qufas-ai.vercel.app/v1/images/generations \2 -H "Authorization: Bearer $QUFAS_API_KEY" \3 -H "Idempotency-Key: image-order-4821-v1" \4 -H "Content-Type: application/json" \5 -d '{"model":"<enabled-image-model-id>","prompt":"A blue robot","response_format":"url"}'Correct usage
- Generate one stable value for one logical operation.
- Reuse it after a timeout or uncertain transport failure for that same operation.
- Do not reuse a value across unrelated prompts, users, or operations.
- Keep the request payload consistent when reusing the key.
Request ID is different
x-qufas-request-id is the canonical trace identifier generated for each gateway request. A retry may have a new request ID while retaining the same idempotency key.
Was this page helpful?