generated: '2026-08-14'
method: searched
source: https://developers.cloudtalk.io/swagger.json
docs: https://developers.cloudtalk.io/
note: >-
  Limits and the exact runtime response headers are published by CloudTalk in the "Usage" section of
  info.description in their own OpenAPI 3.0.1 v1.7 document, fetched verbatim from
  https://developers.cloudtalk.io/swagger.json (HTTP 200, application/json) on 2026-08-14. The same
  text renders on the developer reference at https://developers.cloudtalk.io/.
limit_count: 1
limits:
- scope: per-company
  window: 1 minute
  limit: 60
  unit: requests
  burst: null
  applies_to: all operations across https://my.cloudtalk.io/api
  note: >-
    "the default rate limit is 60 operations per minute per company" — an operation is a read or an
    update. Limits are enforced per company (account), not per API key, so multiple integrations
    sharing one CloudTalk account contend for the same 60/min budget.
  increase: >-
    "If you need to make more requests, please contact our support with detailed explanation of your
    use case." No self-serve limit increase and no published tier-based ceilings.
exhaustion:
  status_code: 429
  status_text: Too Many Requests
  retry_after_header: false
  note: >-
    CloudTalk returns 429 and the three X-CloudTalkAPI-* headers below. It does NOT return a standard
    Retry-After header, and does not implement the IETF draft RateLimit-* headers. A client must
    compute the wait itself from X-CloudTalkAPI-ResetTime.
response_headers:
- name: X-CloudTalkAPI-Limit
  description: Maximum number of API requests allowed in the current time window.
  standard: vendor-specific
- name: X-CloudTalkAPI-Remaining
  description: Number of API requests left in the current window.
  standard: vendor-specific
- name: X-CloudTalkAPI-ResetTime
  description: Time when the rate limit window will be reset, as a Unix timestamp.
  standard: vendor-specific
standard_headers:
  ietf_ratelimit: false
  retry_after: false
  note: >-
    No `RateLimit-Limit`/`RateLimit-Remaining`/`RateLimit-Reset` (IETF draft) and no `Retry-After`.
    Agent clients must special-case the X-CloudTalkAPI-* triple.
agent_guidance: >-
  Read X-CloudTalkAPI-Remaining on every response and back off before it reaches zero; on a 429, sleep
  until the Unix timestamp in X-CloudTalkAPI-ResetTime rather than using a fixed backoff. Because the
  budget is per company, an agent sharing an account with human users or other integrations should
  reserve headroom rather than consuming the full 60/min.
related:
- conventions/cloudtalk-conventions.yml
- errors/cloudtalk-problem-types.yml