generated: '2026-07-18'
method: searched
source: https://docs.armada.ai/bridge/api
docs: https://docs.armada.ai/bridge/tenant-guide/ai-studio/model-as-a-service/api-key
# Cross-cutting request/response semantics for the Armada Bridge platform,
# derived from the Orchestrator OpenAPI and the published tenant/admin docs.
# Cross-links: authentication/, errors/, lifecycle/.
authentication:
  style: bearer-token
  detail: >-
    Orchestrator API requires a Bearer JWT (Authorization: Bearer <token>) issued by the
    platform auth middleware. The Model-as-a-Service inference API uses an API key presented
    as a Bearer token (Authorization: Bearer $API_KEY), generated and revoked in AI Studio
    Settings and displayed once at creation. Console/UI sign-in is OIDC via Keycloak.
  ref: authentication/armada-authentication.yml
error_envelope:
  format: http-status
  media_type: application/json
  note: Standard HTTP status codes (400/401/403/404/409/500) with a human-readable description.
  ref: errors/armada-problem-types.yml
versioning:
  scheme: path-relative
  api_version: '1.0.0'
  detail: >-
    Orchestrator API is v1; servers[] is declared relative ("/") and must be prefixed with
    the deployment base URL, since Bridge is deployed per tenant/edge site rather than as a
    single shared public host. Documentation is independently versioned (current line 5.3.0).
  ref: lifecycle/armada-lifecycle.yml
idempotency:
  supported: false
  note: No Idempotency-Key header or parameter is declared in the OpenAPI or documented.
pagination:
  documented: false
  note: No cursor/offset pagination scheme is declared in the OpenAPI or documented.
rate_limiting:
  documented: false
compatibility:
  model_inference: >-
    The Bridge AI Studio Model-as-a-Service endpoints are OpenAI-compatible
    (POST /v1/chat/completions with model/messages/temperature/max_tokens),
    served from a per-deployment host shown on the Playground page.