generated: '2026-09-05'
method: searched
source: https://developer.opus.com/api-reference/v1-workflow-generation/generate-headless
docs:
- https://developer.opus.com/api-reference/v1-workflow-generation/workflow-generation-introduction
- https://developer.opus.com/api-reference/v1-integration/integration-introduction
- https://developer.opus.com/mcp-setup
limit_count: 3
note: >-
  AAICO publishes concurrency limits with real numbers, but does NOT publish the numeric rate limit
  on the API key itself, and publishes no rate-limit response headers on the platform API. The
  documentation-MCP server is the only surface with a published request-rate number.
rate_limits:
- id: active-runs-per-workspace
  scope: per-workspace
  kind: concurrency
  limit: 20
  unit: active generation runs
  window: concurrent
  status_on_exhaustion: 429
  response_body_fields: [scope, limit, active]
  headers: none documented
  api: Opus Platform API
  source: https://developer.opus.com/api-reference/v1-workflow-generation/generate-headless
- id: active-runs-per-organization
  scope: per-organization
  kind: concurrency
  limit: 100
  unit: active generation runs
  window: concurrent
  status_on_exhaustion: 429
  response_body_fields: [scope, limit, active]
  headers: none documented
  api: Opus Platform API
  source: https://developer.opus.com/api-reference/v1-workflow-generation/generate-headless
- id: mcp-docs-per-ip
  scope: per-user (IP)
  kind: request-rate
  limit: 200
  unit: requests
  window: 1h
  status_on_exhaustion: not documented
  headers: none documented
  api: Opus Docs MCP Server
  source: https://developer.opus.com/mcp-setup
undocumented_limits:
- id: api-key-burst
  scope: per-key
  kind: request-rate
  limit: null
  window: null
  status_on_exhaustion: 429
  note: >-
    The docs state twice that "keys are rate limited — sustained bursts return 429, independent of
    the active-run limits", but never publish the number or the window. An integrator cannot size
    a client against it.
  source: https://developer.opus.com/api-reference/v1-workflow-generation/workflow-generation-introduction
per_workflow:
  concurrent_generation_runs: 1
  status_on_exhaustion: 409
  note: A workflow may have only one generation run in progress; a second returns 409 Conflict.
  source: https://developer.opus.com/api-reference/v1-workflow-generation/generate-headless
other_limits:
- id: file-size
  limit: 10
  unit: MB
  scope: per uploaded file
  source: https://developer.opus.com/api-reference/introduction
- id: presigned-url-expiry
  min: 60
  max: 86400
  unit: seconds
  scope: per presigned download URL
  source: https://developer.opus.com/opus-sdk/overview
response_headers:
  published: []
  x_ratelimit: false
  ratelimit_rfc9331: false
  retry_after: false
  gap: >-
    No X-RateLimit-*, RateLimit-* or Retry-After header is documented on any endpoint. The only
    runtime signal is the 429 body's {scope, limit, active}, and only for the active-run limit.