Rate limits

Requests per minute, per key

• Every REST call counts once, and so does every MCP tool call, whether it comes with an API key or an OAuth token. The limit follows the workspace's plan the moment it changes.
• Starting projects and carousels (POST /projects, POST /uploads, POST /clips/{id}/carousels and POST /carousels/{id}/posts) has a second limit of 10 a minute per workspace, across all its keys, because each one holds hours or runs a model.
• Windows are fixed 60-second windows counted in the database, so the limit holds across every server. A burst can reach up to twice the limit across a window edge.
• Need more? Email support@reeltopus.com with your workspace and expected volume.

Headers and 429s

Every response says where you stand:

RateLimit-Limit: 300
RateLimit-Remaining: 287
RateLimit-Reset: 41          # seconds until the window resets

Over the limit, the API answers 429 with a Retry-After header and the same wait in the body:

HTTP/1.1 429 Too Many Requests
Retry-After: 41

{ "error": { "code": "rate_limited", "message": "Too many requests. Try again in 41 s.", "retryAfterSeconds": 41 } }

The SDKs wait for Retry-After and retry on their own, with the same Idempotency-Key, so a retried write never runs twice.

Other limits

Processing spends the workspace's hours like an upload in the app. A request that needs more hours than the workspace has left answers 402 with code not_enough_hours.