Rate limits protect the platform and ensure fair capacity across accounts. Limits apply per API key and environment (live vs test).

​
Hosted account limits (default)

Contact support to raise limits on enterprise plans.

​
HTTP response when limited

Honor Retry-After and use exponential backoff with jitter.

​
Session concurrency

Even within account rate limits, the orchestrator serializes requests per session_id. If a second request arrives while the first holds the session lock:
HTTP 429 — this is per-session, not a global account cap. Use distinct session IDs for parallel independent scrapes.

​
Best practices

  • Reuse session_id for multi-step flows instead of opening many parallel sessions to the same site
  • Set tag to identify noisy jobs in logs and support requests
  • Batch large workloads with a job queue in your application instead of opening hundreds of parallel sync connections
  • Cache stable pages client-side when polling on an interval