rate_limited

What it means

The per-minute burst limit for this key was exceeded (60/min on Free, 300/min on Pro). This is about request PACING, not your monthly quota.

Common causes

• A burst of parallel requests inside one minute window.
• A retry loop without backoff.

How to fix it

Honor the `Retry-After` header (seconds until the window rolls), add exponential backoff, and spread bulk work across minutes. The `ratelimit` response headers show the live window state on every response.

Should you retry?

Yes — after `Retry-After` seconds. The official SDK does this automatically.

The response shape

Every /v1 error is an RFC 9457 application/problem+json document: type links to this page, code is the stable value to branch on (never parse title or detail), and request_id identifies the request to support. Validation failures add an errors array of per-field JSON pointers.

API overview · API reference · All error codes:

• missing_credentials — HTTP 401, Missing credentials
• invalid_api_key — HTTP 401, Invalid API key
• invalid_token — HTTP 401, Invalid access token
• unsupported_credential_type — HTTP 403, Credential type not usable on this surface
• insufficient_scope — HTTP 403, Insufficient scope
• feature_not_enabled — HTTP 403, Feature not enabled on this plan
• validation_failed — HTTP 400, Request validation failed
• resource_not_found — HTTP 404, Resource not found
• quota_exhausted — HTTP 429, Plan quota exhausted
• idempotency_key_reused — HTTP 422, Idempotency-Key reused with a different request
• idempotency_in_progress — HTTP 409, A request with this Idempotency-Key is still in progress
• internal_error — HTTP 500, Internal server error