generated: '2026-09-09'
method: searched
source: https://comunicate.top/ro/documentatie-api
limit_count: 2
limits:
  - scope: per-key
    window: 1m
    limit: set per key at issue time (written on the key when issued; not one global number)
    status_on_exhaustion: 429
    headers: [Retry-After]
    note: >-
      Limits are per key, not per account, so a runaway integration does not block the account's
      other keys. 429 with Retry-After is documented as the only case where retrying makes sense
      (vs 403, which is not retryable).
  - scope: per-organization
    window: 1h
    limit: 60
    applies_to: AI routes — article writing, document import, metadata completion
    status_on_exhaustion: 429
    headers: [Retry-After]
    note: >-
      Applies regardless of how many keys the organization holds; these routes consume the
      platform's own model quotas.
observed:
  - url: https://app.comunicate.top/api/v1/public/stats
    status: 200
    rate_limit_headers: none
    note: >-
      Public keyless routes advertise "no rate-limit headers to negotiate" in the OpenAPI info
      description; a live unauthenticated response carried no X-RateLimit-*/RateLimit-* headers.