generated: '2026-08-17'
method: searched
source: https://www.edgee.ai/docs/api-reference
also:
- https://www.edgee.ai/docs/api-reference/errors
- https://www.edgee.ai/docs/features/usage-limits
note: 'Edgee publishes NO numeric rate limit. The API reference says only that "Edgee has its own
  rate limit technology to prevent abuse and ensure service stability" and that exceeding it returns
  429 — no requests-per-second, per-minute or per-day figure, no per-endpoint table, and no
  RateLimit-* / X-RateLimit-* response headers in the docs or in the OpenAPI. What Edgee DOES publish
  in detail is a spend-limit system (usage limits and budgets), which throttles on dollars rather
  than on request rate and shares the same 429 status. Both are recorded below and kept distinct:
  limit_count counts published numeric REQUEST-rate limits, and that number is 0.'
limit_count: 0
rate_limits: []
response_signaling:
  status_on_exhaustion: 429
  error_code: usage_limit_exceeded
  error_type: rate_limit_error
  headers:
  - name: Retry-After
    published: conditional
    note: 'Docs instruct clients to "wait for the time specified in the Retry-After header (if
      present)". Presence is not guaranteed and the header is not declared in the OpenAPI.'
  - name: RateLimit-*
    published: false
  - name: X-RateLimit-*
    published: false
  guidance: Implement exponential backoff on 429; reduce request rate.
  source: https://www.edgee.ai/docs/api-reference/errors
spend_limits:
  documented: true
  source: https://www.edgee.ai/docs/features/usage-limits
  description: 'Spend caps, not request-rate caps. Enforced per API key, per member, or per squad,
    over a daily, monthly or lifetime window. Exceeding a cap returns the same 429 with
    usage_limit_exceeded.'
  scopes: [api-key, member, squad]
  windows: [daily, monthly, lifetime]
  unit: USD credits
  alerting:
    source: https://www.edgee.ai/docs/features/alerts
    thresholds: ['50%', '80%', '100%']
    alert_types: [api-key-budget, member-budget, squad-budget, tag-spend, remaining-credits]
    channels: [email, slack]
    tag_spend_windows: [1h, 3h, 6h, 12h, 24h]
probe:
  url: https://api.edgee.ai/v1/models
  method: GET
  http_status: 200
  fetched: '2026-08-17'
  observed_headers: [date, content-type, content-length]
  finding: 'An unauthenticated 200 from the live model catalog returned no rate-limit headers of any
    kind, which is consistent with the docs.'