Errors
AdaptiqAI returns explicit HTTP errors instead of silently dropping required Native Memory or crossing an isolation boundary.
| Status | Type | Meaning |
|---|---|---|
400 | invalid_request_error | The request does not match the OpenAI-compatible schema. |
401 | authentication_error | The API key is missing, invalid or revoked. |
403 | permission_error | The key cannot access the requested workspace or model. |
404 | not_found_error | The requested model or resource is not available. |
409 | conflict_error | The requested memory operation conflicts with current state. |
429 | rate_limit_error | The workspace has reached an active rate or concurrency limit. |
503 | service_unavailable | The provider cannot safely execute the request right now. |
Retry deliberately
Retry 429 and transient 503 responses with exponential backoff and jitter. Do not retry authentication, permission or schema errors without changing the request.