generated: '2026-08-26'
method: searched
source: >-
  https://api-docs.tgl.ai/ (password-protected) plus observed response headers on
  https://api.reflexivity.com/alfred/v1
note: >-
  Reflexivity publishes no rate limits. The API reference that would carry them is password-protected,
  and an unauthenticated request to the production API host returns 401 with no RateLimit-*,
  X-RateLimit-* or Retry-After header of any kind — so neither the documented number nor the runtime
  signal is available to an agent before it authenticates. An honest zero.
  The one runtime signal the gateway does emit anonymously is `x-request-id` (a UUID) and an
  `access-control-expose-headers: Conversation-Id`; those are traceability, not throttling.
limit_count: 0
limits: []
response_headers:
  rate_limit_headers_observed: []
  retry_after_observed: false
  exhaustion_status_code: null
  other_headers_observed:
  - x-request-id
  - access-control-expose-headers: Conversation-Id
  - x-envoy-upstream-service-time
observed_on:
- url: https://api.reflexivity.com/alfred/v1
  method: GET
  status: 401
  note: unauthenticated probe; no throttling headers present