Rate limits and 429s

Request limits are shared across your account's API keys, dashboard, and MCP connections: 10 per minute on Free, 60 on Basic, 300 on Pro, and 900 on Mega. Scale starts at 900 per minute, with custom increases available. Creating more keys does not increase the allowance. A lower limit on a particular key still applies.

Concurrent work is also limited by plan. Batches share that capacity with individual requests. See the rate-limit reference for the full table.

Headers

Responses from limited endpoints include the current allowance. For example, on Basic:

X-RateLimit-Limit: 60
X-RateLimit-Remaining: 58
X-RateLimit-Reset: 1717619400

When you exceed it

You get 429 Too Many Requests with a Retry-After header in seconds. Wait that long before retrying. Rejected attempts do not extend the request window, but repeated immediate retries cannot make capacity available sooner.

Job-status polls have a separate allowance of twice your account's request limit, with a minimum of 120 per minute. They do not consume extraction slots or credits.

Rate limits are not credits

A 429 costs no credit. It means you are sending too fast, not that you have run out.

Limits are per account rather than per key, and are listed with the current plans.