generated: '2026-08-23'
method: searched
source: https://partner-api-docs.myalex.com/
docs: https://partner-api-docs.myalex.com/
limit_count: 0
summary: >-
  Rate limiting exists and is acknowledged in the docs, but no number is published. The V1 API
  Reference states that requesting an access token with each API call "will cause rate limiting
  errors" — confirming a limit is enforced on the Auth0 token endpoint — without stating the
  threshold, the window, the scope it applies to, or the status code and headers returned on
  exhaustion. No limit of any kind is published for the three Partner Integration API endpoints
  themselves.
rate_limits: []
documented_constraints:
- scope: token endpoint
  endpoint: /oauth/token (login.myalex.com and login.jellydevs.com)
  window: not published
  limit: not published
  burst: not published
  statement: >-
    "If an access token is requested with each API call, rate limiting errors will occur."
  source_section: Authentication > Guardrails
- scope: token lifetime
  value: 86400 seconds (24 hours)
  statement: >-
    Documented expires_in is 86400. The published guardrail is to cache the token and re-mint every
    24 hours or on a 401 — this token-reuse requirement is effectively the published rate-limit
    mitigation.
response_headers:
  published: false
  observed: false
  headers: []
  detail: >-
    No RateLimit-Limit / RateLimit-Remaining / RateLimit-Reset (RFC 9331 style), no legacy
    X-RateLimit-* set, and no Retry-After is documented anywhere in the reference. The HTTP status
    table published in the appendix lists 200, 201, 400, 401, 403, 404, 405, 500 and 503 — 429 Too
    Many Requests is not among them, so the status returned on exhaustion is unstated.
exhaustion_status_code: not published
probe_note: >-
  Not probed live. Both the token endpoints and the three API endpoints require credentials issued by
  Jellyvision during partner onboarding, so no unauthenticated request can observe a runtime
  rate-limit header without provisioned access.
agent_impact: >-
  An agent integrating this API cannot compute a safe call rate or implement header-driven backoff.
  The only actionable runtime signal is the documented 401-then-re-mint guardrail; everything else
  requires retry-with-backoff on 503 and escalation using the response requestId.