generated: '2026-08-17'
method: searched
source: >-
  https://api.lengow.io/docs/ (the published OpenAPI 3.0.3 — GET /rate-limits operation and the
  components.responses.TooManyRequests definition) and https://docs.lengow.io/ (per-endpoint limit notes)
docs: https://api.lengow.io/docs/
limit_count: 2
note: >-
  Lengow does NOT publish a global numeric quota. It publishes something better for an agent: an
  authenticated introspection endpoint, GET /rate-limits, that returns every API route available to the
  calling login together with that route's configured fixed-window quota, the remaining credit, and the
  seconds left in the window. The only hard number stated in public prose is the 10 req/min cap on the
  v1.0 report export. The 429 example in the spec ("Request limitation: 500 reached on /v3.0/orders")
  shows the envelope, not a guaranteed 500/window contract, so it is recorded as an example below rather
  than as a published limit.
model: fixed-window, per-route, per-login (Redis bucket)
introspection:
  endpoint: GET https://api.lengow.io/rate-limits
  auth_required: true
  returns:
  - route
  - configured quota for the route
  - remaining_credit (current Redis bucket value, clamped to zero after exhaustion)
  - remaining_time (seconds until the fixed window expires)
  null_semantics: remaining_credit and remaining_time are null when rate limiting is disabled for a route
limits:
- scope: per-endpoint
  endpoint: GET /v1.0/report/export
  window: 1 minute
  limit: 10
  source: https://docs.lengow.io/
- scope: per-route-per-login
  endpoint: all routes
  window: fixed window (length returned per route by GET /rate-limits)
  limit: configured per route; not published as a static number
  source: https://api.lengow.io/docs/ (GET /rate-limits)
exhaustion:
  status: 429
  response_headers:
  - name: Retry-After
    description: Seconds until the current fixed window expires
    type: integer
    minimum: 0
  body_example: '{"error": {"code": 429, "message": "Request limitation: 500 reached on /v3.0/orders"}}'
  note: >-
    Retry-After is the ONLY rate-limit response header the spec declares. No X-RateLimit-* or RFC 9331
    RateLimit-* headers are documented on successful responses, so a client cannot see remaining budget
    without an extra call to GET /rate-limits.
pagination_interaction:
  note: >-
    Orders list responses are capped at 100 orders per page (page / page_size), and the v1.0 catalogue
    endpoints paginate with page / limit — the practical throughput ceiling for a full catalogue sync.