generated: '2026-08-11'
method: searched
source: https://proofdraw.com/openapi.json
docs: https://proofdraw.com/api
limit_count: 0
note: >-
  ProofDraw publishes a rate-limit SIGNAL but no rate-limit NUMBERS. The OpenAPI info.description states:
  "Rate limits: per API key. `X-RateLimit-Limit` and `X-RateLimit-Remaining` headers are returned on every
  response." No quota value, window, burst allowance, per-endpoint variation, or exhaustion status code is
  published anywhere on the public surface. The docs defer the detail to docs/api-spec.md, which 404s.
  An agent can read remaining budget at runtime but cannot plan against a documented ceiling.
headers:
  - name: X-RateLimit-Limit
    documented: true
    observed: false
  - name: X-RateLimit-Remaining
    documented: true
    observed: false
  - name: X-RateLimit-Reset
    documented: false
    observed: false
  - name: Retry-After
    documented: false
    observed: false
exhaustion:
  status_code: null
  error_code: rate_limited
  note: >-
    `rate_limited` appears in the Error `code` enum of the published OpenAPI, but no operation declares a
    429 (or any) response for it. The status an agent should branch on is genuinely undocumented.
limits: []
probe:
  - url: https://proofdraw.com/api/health
    status: 200
    date: '2026-08-11'
    rate_limit_headers_present: false
    observed_headers:
      - date
      - content-type
      - cache-control
      - access-control-allow-origin
      - cf-cache-status
    note: >-
      The one unauthenticated JSON endpoint returns NO X-RateLimit-* headers, contradicting "returned on
      every response". The headers are most likely emitted only on authenticated (per-key) requests, which
      cannot be verified without credentials — but as published, the claim is not observable anonymously.
tier_quotas:
  note: >-
    Plan caps are a separate mechanism from rate limiting and are enforced as 403 tier_limit_exceeded /
    422 entry_limit_exceeded rather than as throttling. See plans/proofdraw-plans-pricing.yml.