specification: API Commons Rate Limits
specificationVersion: '0.1'
provider: Lumos
providerId: lumos
generated: '2026-08-29'
method: probed
source: >-
  Live unauthenticated response headers observed on https://api.lumos.com/info on 2026-08-29, plus the
  429 semantics declared in https://api.lumos.com/openapi.json
created: '2026-05-04'
modified: '2026-08-29'
description: >-
  Lumos returns runtime rate-limit headers on every response, but publishes no rate-limit table in its
  documentation. The numbers below are what the edge actually returned to an anonymous caller; the
  authenticated per-key allowance is not documented anywhere on developers.lumos.com.
  This file replaces a 2026-05-04 bulk-sweep scaffold whose tier limits were invented defaults.
headers:
  limit: x-ratelimit-limit
  remaining: x-ratelimit-remaining
  reset: x-ratelimit-reset
  retryAfter: retry-after
  policy: null
  reset_format: unix epoch seconds (float)
  observed: true
responseCodes:
  throttled: 429
  quotaExceeded: 429
  serviceUnavailable: 503
limit_count: 1
limits:
- name: Unauthenticated edge limit (observed)
  scope: per-ip
  window: not stated
  limit: 10
  burst: null
  headers_returned:
    x-ratelimit-limit: '10'
    x-ratelimit-remaining: '9'
    x-ratelimit-reset: '1788007986.504639'
    retry-after: '1'
  method: probed
  evidence:
    url: https://api.lumos.com/info
    http_status: 200
    probed: '2026-08-29'
  note: >-
    Observed on an anonymous GET. This is the edge allowance an unauthenticated caller sees; it is NOT
    the documented API-key allowance, which Lumos does not publish.
documented_limits:
  count: 0
  note: >-
    No rate-limit reference page exists on developers.lumos.com — the llms.txt index of 223 pages
    contains no rate-limit, quota, or throttling entry. An honest zero.
operation_specific:
- operationId: reconnectApp
  path: PUT /apps/{app_id}
  status: 429
  condition: >-
    "Too many requests — a sync is already in flight for this app. Includes a `Retry-After` header
    (seconds)." This is a concurrency lock expressed as a 429, not a rate limit.
  source: openapi/lumos-openapi.json
- operationId: createKnowledgeContext
  path: POST /v1/knowledge/context
  status: 429
  condition: 429 declared on all five Knowledge Context operations.
  source: openapi/lumos-openapi.json