generated: '2026-08-13'
method: searched
source: https://sms.cresclab.com/developers.html + https://sms.cresclab.com/install.html + openapi/_original/crescendo-lab-maacgo-openapi.yaml
api: MAAC Go API
base_url: https://sms.cresclab.com/api
limit_count: 2
summary: >-
  MAAC Go publishes account-age-tiered send limits rather than per-endpoint
  quotas. New accounts sit under a probation cap for their first 24 hours, then
  move to the established-account hourly cap. Exhaustion returns HTTP 429 with
  the machine code `rate_limited`.
discrepancy:
  note: >-
    The two published pages disagree on the new-account number and this is
    recorded rather than resolved. The developer page (developers.html,
    "Rate limit 清楚" panel) states 新帳號 24h 內 200 則 (200 messages in 24h);
    the install manual (install.html, Q5 troubleshooting table) states
    新帳號（< 24h）24h 內 ≤ 2,000 通 (2,000 messages in 24h). Both pages agree on
    the established-account figure of 1,000/hour. An integrator should assume
    the lower bound.
  sources:
  - https://sms.cresclab.com/developers.html
  - https://sms.cresclab.com/install.html
limits:
- name: New account probation cap
  scope: per-account
  applies_to: accounts less than 24 hours old
  window: 24h
  limit: 200
  limit_upper_published: 2000
  unit: messages
  note: Lower bound from developers.html; install.html publishes 2,000. See discrepancy above.
- name: Established account cap
  scope: per-account
  applies_to: accounts 24 hours old or more
  window: 1h
  limit: 1000
  unit: messages
  note: Consistent across both published pages.
exhaustion:
  status_code: 429
  error_code: rate_limited
  body: >-
    Standard MAAC Go error envelope ({error, hint, issues?}) carrying limit/used
    detail. Documented in the OpenAPI info block as "Rate-limit errors return
    429 rate_limited with limit/used details and retry guidance when available."
  operations_declaring_429:
  - sendSms
  - createBroadcast
response_headers:
  documented:
  - name: retry_after
    note: >-
      developers.html states "超量回 429 帶 retry_after" (429 carries retry_after).
      The docs render it lowercase; the OpenAPI does not declare response headers
      for any operation, so the exact casing (retry_after vs Retry-After) is not
      confirmed in a machine-readable contract.
  standard_ratelimit_headers:
    present: false
    note: >-
      No RateLimit-* (RFC 9331 draft) or X-RateLimit-* headers are documented and
      none are declared in the OpenAPI. An agent cannot read remaining quota
      before exhaustion; it must handle the 429.
  verification:
    method: not observed on a live response
    reason: >-
      Every /api endpoint requires a bearer key (unauthenticated calls return
      401 before any rate-limit header would be emitted), and robots.txt
      disallows /api/, so headers were not probed on a live 429.
remediation:
  documented: Wait out the cooldown, or contact info@cresclab.com to raise the cap.
  source: https://sms.cresclab.com/install.html
other_bounds:
  note: Non-rate limits that bound throughput, harvested from the OpenAPI.
  broadcast_inline_threshold: 30 recipients — batches of 30 or fewer dispatch inline; more than 30 queue asynchronously
  contacts_list_limit: page + limit, limit maximum 500, default 50
  sms_list_limit: limit maximum 200, default 50
  wallet: >-
    A prepaid wallet is the real throughput ceiling in practice — a send that
    exceeds the balance returns HTTP 402 insufficient_balance with a topup_url,
    independently of the rate limit.
cross_links:
  errors: errors/crescendo-lab-problem-types.yml
  conventions: conventions/crescendo-lab-conventions.yml
  plans: plans/crescendo-lab-plans-pricing.yml