Rate limits

Limits are per account, not per key. The capture limits count dashboard captures alongside API calls, because the cap is on what your account asks the platform to do, not on which door the request came through.

What bounds usage

• 60 captures per minute, per account — that is POST /api/v1/screenshots.
• 5 captures running at once (queued or processing), per account.
• 300 reads per minute, per account, shared between GET and DELETE on a job. The 5-second polling interval Check job status recommends stays well inside this, even with every capture you are allowed to run at once being polled at that rate.
• Your credit balance. One completed screenshot is one credit — see Credits.
• Capture capacity. When the queue is saturated the service answers 503 regardless of how much of your own allowance you have left.

Backing off

A throttled request answers 429 with a Retry-After header in seconds. Wait at least that long, then retry with exponential backoff. That header is the only rate-limit signal the service sends today — there are no X-RateLimit-* headers to read.

HTTP/1.1 429 Too Many Requests
Retry-After: 12

These are deliberately conservative opening numbers, not tuned ones. If you are planning volume that would need more, talk to us first.