Metered API-key requests share the organization’s billing-period allowance. Creating more keys does not create more allowance. Signed-in dashboard requests do not consume this API quota.
Read the quota headers
Metered responses expose X-Quota-Limit, X-Quota-Remaining and X-Quota-Reset. The reset value is a Unix timestamp in seconds.
curl --include "$CHARTEROPS_API_URL/v1/aircraft" \
--header "Authorization: Bearer $CHARTEROPS_API_KEY"See plans and API limits for the monthly allowances, per-key rates and optional request packs. Packs add monthly capacity; there are no automatic overage charges.
Each key has a per-minute limit
Every key may make its plan’s requests_per_minute in one UTC minute. The pricing page lists each plan’s current rate. The limit is each key’s own, so one busy integration cannot slow another. Metered responses expose X-RateLimit-Limit, X-RateLimit-Remaining and X-RateLimit-Reset, a Unix timestamp in seconds for the start of the next minute.
A key past its limit returns 429 with the problem type rate_limited, a reset_at timestamp and Retry-After in seconds. The refused request does not count against the billing-period allowance. Wait for Retry-After, then continue; spread a batch across the minute rather than sending it at once.
When the allowance runs out
An exhausted allowance returns 429 with the problem type quota_exceeded, a reset_at timestamp and an upgrade_url. Respect Retry-After, expressed in seconds. Waiting for the next minute will not reset a billing-period quota.
Tell the two 429 responses apart by their type: rate_limited clears within a minute, quota_exceeded clears when the period resets.
Public and authentication routes share a limit of 60 requests per IP address per UTC minute. This includes sign-in, organization creation, invitation previews, plans and API documentation. Token renewal has its own allowance of 60 a minute per address, so it is never refused because of other traffic. One address may also ask for at most thirty sign-in codes an hour (a new email’s first code creates its account) and create at most ten organizations an hour. One email receives at most five codes an hour, and a code stops working after five wrong guesses. Signed-in dashboard sessions are limited to 300 requests a minute per person, far above what the dashboard asks for. A rate_limited response includes reset_at and Retry-After; wait until that window resets before trying again.
What counts
Requests admitted to metered routes count, including admitted 4xx failures. 429 and 5xx responses are refunded. Some operations, including posting and editing empty legs, are unmetered. The API reference describes individual operations.
See what your keys did
The dashboard’s API page has a Logs tab. It lists every request your keys made in the last 30 days with its route, status, response time and request ID, and shows the success rate and response time for the period and key you choose. Open a failed request to read the message it was answered with and, for a 422, each field that needs correcting. Filter by key or status, or paste a request ID to find one call.