Limits and credits
Credits and prices
Runs started through the API cost the same credits at the same prices as runs started in the dashboard. They draw from your workspace's balance.
Caps per key
You can give a key a daily, weekly or monthly credit cap. All three are optional.
- The daily cap resets at 00:00 UTC.
- The weekly cap resets on Monday at 00:00 UTC.
- The monthly cap resets on the first of the month at 00:00 UTC.
A spend counts in the window in which it was reserved. ReelBolt checks the caps before a run starts and again before every step. A run that reaches a cap stops at that step, and earlier steps keep their charge. A step that costs a little more than its estimate can push the next check over the cap.
Caps apply only when billing is on for your workspace. When billing is off, no credits are reserved and caps are not enforced.
When a cap or the balance blocks a run
A start that would exceed a cap answers 402:
{
"error": {
"code": "API_KEY_BUDGET_EXCEEDED",
"message": "...",
"window": "daily",
"capCredits": 500,
"usedCredits": 480,
"resetsAt": "2026-11-03T00:00:00Z"
}
}
window is the cap that blocked the run. Wait until resetsAt, raise the cap, or start a smaller run.
Running out of workspace credits answers 402 with INSUFFICIENT_CREDITS instead.
Rate limits
Each key can make 300 requests a minute, and 60 of them can be writes (POST, PUT, PATCH and DELETE). Authenticated responses that reach the rate limiter
carry:
| Header | Meaning |
|---|---|
RateLimit-Limit | Requests allowed in the window |
RateLimit-Remaining | Requests left in the window |
RateLimit-Reset | Seconds until the window resets |
401, 402 and 403 refusals, and a 429 from the edge limit, may not carry these headers.
Over the limit, the answer is 429 with error.code rate_limited, a Retry-After header in seconds and
error.resetsAt as a UTC time. Wait that long before retrying. Starting runs has a separate hourly limit: a start over it is queued, and a resume or retry answers 429 RUN_RATE_LIMITED.
See Runs.
Idempotency keys
Send an Idempotency-Key header on a POST so that a retry after a network error does not repeat the work.
ReelBolt remembers the key for 24 hours for each API key.
- The same key with the same body returns the stored answer.
- The same key with a different body returns
422idempotency_key_reused. - The same key while the first request is still running returns
409idempotency_in_progress. - A key that is not a valid value returns
invalid_idempotency_key.
Use a new random value, such as a UUID, for each new action.