generated: '2026-08-23'
method: searched
source: >-
  openapi/kortext-labs-api-openapi.json, live unauthenticated responses from
  https://api-demo.labs.kortext.com and https://vle.kortext.com, and a search of every
  Kortext-published surface for a limits/throttling statement, 2026-08-23.
limit_count: 0
note: >-
  Kortext documents no rate limits anywhere. The published contract declares no 429 response on any
  of its 67 operations and no header parameters at all. Live unauthenticated responses from
  api-demo.labs.kortext.com carry no X-RateLimit-*, RateLimit-* or Retry-After header. Because
  there is no developer documentation, there is nothing to search — this zero is measured, not
  assumed. Note that this API fronts expensive generative operations (podcast synthesis, quiz
  generation, document indexing), so the absence of a documented limit is a real integration
  hazard rather than a formality.
limits: []
response_headers:
  observed: []
  declared: []
  probe:
  - url: https://api-demo.labs.kortext.com/health/live
    status: 200
    rate_limit_headers: none
  - url: https://api-demo.labs.kortext.com/podcast/v1/podcasts
    status: 403
    body: '{"detail":"Not authenticated"}'
    rate_limit_headers: none
status_on_exhaustion: undocumented
declared_429_operations: 0
edge_behaviour:
  gateway: Azure Application Gateway WAF (Prevention mode)
  evidence: >-
    Every /swagger* path on every Kortext host returns HTTP 403 with a Kortext-branded error page
    carrying the ops marker "Request blocked by the Application Gateway WAF (Prevention mode)" and
    reference KX-403. This is a WAF path block, not a rate limit, but it is the only throttling or
    blocking behaviour Kortext exposes to an unauthenticated caller.