- name: Public API metadata, pricing and OAuth
  type: Endpoint
  limit: 12
  metric: request
  timeframe: minute
  description: Public metadata, pricing and OAuth routes. Shared per source IP at each gateway replica, not a per-account
    allowance. Additional capacity controls can apply. Honor Retry-After when present; otherwise use bounded backoff.
  domains:
  - api.machinelibrary.ai
  url: https://machinelibrary.ai/docs/api/operations#limits
  source_date: '2026-09-26'
- name: MCP protocol and discovery
  type: Endpoint
  limit: 60
  metric: request
  timeframe: minute
  description: MCP discovery and protocol requests. Shared per source IP at each gateway replica, not a per-account
    allowance. Additional capacity controls can apply. Honor Retry-After when present; otherwise use bounded backoff.
  domains:
  - mcp.machinelibrary.ai
  url: https://machinelibrary.ai/docs/api/operations#limits
  source_date: '2026-09-26'
- name: Recognition requests
  type: Endpoint
  limit: 30
  metric: request
  timeframe: minute
  description: Recognition routes; concurrent active-job limits also apply. Shared per source IP at each gateway
    replica, not a per-account allowance. Additional capacity controls can apply. Honor Retry-After when present;
    otherwise use bounded backoff.
  domains:
  - api.machinelibrary.ai
  url: https://machinelibrary.ai/docs/api/operations#limits
  source_date: '2026-09-26'