generated: '2026-07-19'
method: searched
source: openapi/gc-ai-openapi-original.yml
docs: https://docs.gc.ai/api-reference/introduction
summary: >-
  Cross-cutting request/response semantics for the GC AI External API
  (https://app.gc.ai/api/external/v1). REST/JSON, API-key auth, offset
  pagination, an asynchronous job pattern for long-running inference, and a
  standard-header rate-limit signal.
authentication:
  style: api-key
  header: Authorization
  formats:
    org_scoped: gcai_...
    user_scoped: u:gcai_...
  docs: https://docs.gc.ai/api-reference/concepts/api-keys
  ref: authentication/gc-ai-authentication.yml
pagination:
  style: offset
  params: [limit, offset]
  response_field: has_more
  schema: Pagination
  notes: List endpoints accept limit/offset; responses carry a Pagination object with has_more.
async_jobs:
  pattern: 202-accepted-then-poll
  detail: >-
    Long-running requests (chat completions, playbook runs) may return 202 with
    a job_id. Poll GET /jobs/{id} until complete. Chat completions also support
    a bounded wait window that can return 200 inline.
  poll_operation: getAsyncJob
  docs: https://docs.gc.ai/api-reference/concepts/async-jobs
idempotency:
  key_header: false
  notes: >-
    No Idempotency-Key contract is published. Several operations are idempotent
    by design (abortUpload, detachFileFromProject, detachFolderFromProject,
    starProject, unstarProject, and re-sending a numbered upload part
    overwrites it), but there is no client-supplied idempotency key.
versioning:
  scheme: uri-path
  current: v1
  base_url: https://app.gc.ai/api/external/v1
error_envelope:
  media_type: application/json
  shape: '{ error, code?, message?, details? }'
  branch_on: code
  ref: errors/gc-ai-problem-types.yml
rate_limiting:
  status: 429
  code: RATE_LIMITED
  headers: [Retry-After, RateLimit-Limit, RateLimit-Remaining, RateLimit-Reset]
  ref: rate-limits/gc-ai-rate-limits.yml
multi_turn:
  mechanism: Pass the chat_id returned by a completion on the next request to continue the chat.
  docs: https://docs.gc.ai/api-reference/concepts/multi-turn