generated: '2026-08-14'
method: searched
source: https://docs.orthogonal.com/api-reference/introduction
limit_count: 0
description: >-
  Orthogonal acknowledges rate limiting but publishes no numbers. The API
  reference documents a RATE_LIMITED error code ("Too many requests") and the
  provider-published Agent Skill tells agents to prefer batch_use over looping
  "for 20+ calls" because "Orthogonal has rate limits" — but no window, no
  per-key or per-account quota, and no burst allowance is stated anywhere in the
  docs, and no RateLimit response headers are documented.
limits: []
error_signal:
  code: RATE_LIMITED
  http_status: 429
  envelope: '{ "success": false, "error": "Too many requests", "code": "RATE_LIMITED" }'
  source: https://docs.orthogonal.com/api-reference/introduction
  note: >-
    HTTP 429 is inferred from the documented RATE_LIMITED code class; the docs'
    own HTTP status table (400/401/402/404/5xx) does not list 429 explicitly.
headers:
  documented: false
  x_ratelimit: not documented
  ratelimit_rfc9331: not documented
  retry_after: not documented
  note: >-
    No X-RateLimit-*, RateLimit-* (RFC 9331) or Retry-After header is documented,
    so an agent cannot read its remaining budget from a response. It can only
    discover the limit by hitting it.
mitigations:
- id: batch_use
  description: MCP tool running up to 20 API calls in parallel; the documented way to avoid tripping limits.
  source: https://docs.orthogonal.com/mcp/setup
- id: batch_get_details
  description: Fetch parameter info for up to 20 endpoints in one call.
  source: https://docs.orthogonal.com/mcp/setup
- id: parentRequestId
  description: >-
    Pagination continuation token. Follow-on page/offset/cursor calls that pass
    parentRequestId with unchanged non-pagination parameters are charged 0 — a
    cost control rather than a rate control, but it removes the incentive to
    batch pages into a burst.
  source: https://api.orthogonal.com/openapi.json
related:
  conventions: conventions/orthogonal-conventions.yml
  errors: errors/orthogonal-problem-types.yml
  finops: finops/orthogonal-finops.yml
evidence:
- url: https://docs.orthogonal.com/llms-full.txt
  http_status: 200
  finding: 'single rate-limit mention: `RATE_LIMITED` | Too many requests'
- url: https://docs.orthogonal.com/.well-known/agent-skills/orthogonal/skill.md
  http_status: 200
  finding: '"Orthogonal has rate limits. For 20+ calls, use batch_use endpoint instead of looping."'