Rate limits

Requests may be rate limited to keep the platform responsive. A request that exceeds a quota returns 429 Too Many Requests with an error envelope (code rate_limited).

Specific quotas are not yet published while the API is in v0. Build clients to handle 429 gracefully rather than assuming a fixed budget.

Handling 429

• Back off and retry, preferring exponential backoff with jitter.
• If a Retry-After header is present, wait at least that long before retrying.
• Combine retries with idempotency keys on write requests so retries can't double-apply.

Reducing request volume

• Cache responses that don't change often.
• Poll estimates with a sensible interval rather than tight loops; see Estimate a part for the recommended pattern.