> ## Documentation Index > Fetch the complete documentation index at: https://docs.dev.gojinko.com/llms.txt > Use this file to discover all available pages before exploring further. # hotel-ai-search-events > Stream a hotel AI search task's events The task's events as Server-Sent Events, from the start or after `Last-Event-ID`. The stream ends when the task ends its turn, or after 45 seconds: reconnect with the `Last-Event-ID` of the last event received. > [!NOTE] > > The events, their order and how to resume a stream are described in [Waiting for results](https://docs.dev.gojinko.com/ai/hotel-search/waiting-for-results.md#stream-the-events). ## OpenAPI ````yaml api-reference/public-api.yaml GET /v1/hotel_ai_search/tasks/{task_id}/events openapi: 3.0.0 info: title: Jinko Public API version: 0.111.0 description: >- Curated public REST surface for Jinko. Authenticated with jnk_ API keys. See https://docs.gojinko.com for guides. ### Per-end-user attribution On booking calls you may send an optional `X-End-User-Id` request header to attribute the booking to one of your own end users (for per-end-user attribution and rate-limiting). The value is an **opaque, tenant-scoped** identifier that you choose — not a Jinko account id. Omit it to book as the tenant. Values prefixed `user_` or `org_` are rejected. servers: - url: https://api.gojinko.com description: Production - url: https://api.sandbox.gojinko.com description: Sandbox security: [] paths: /v1/hotel_ai_search/tasks/{task_id}/events: get: tags: - Hotels summary: Stream a hotel AI search task's events description: >- The task's events as Server-Sent Events, from the start or after `Last-Event-ID`. The stream ends when the task ends its turn, or after 45 seconds: reconnect with the `Last-Event-ID` of the last event received. parameters: - schema: type: string minLength: 1 maxLength: 100 example: atk_example required: true name: task_id in: path - schema: type: string maxLength: 20 description: The `id` of the last event received. example: '7' required: false name: Last-Event-ID in: header responses: '200': description: The events. content: text/event-stream: schema: type: string description: >- Server-Sent Events, one per task event: `id` is its sequence number (send it back as `Last-Event-ID` to resume), `event` is `task.status` (`status` `submitted` or `working`), `agent.message` (the early reply first), `hotels.page`, or `task.completed` / `task.input_required` / `task.failed` / `task.canceled` / `task.expired` for the status that ends the turn; `data` is JSON and carries `task_id` and `context_id`. The stream ends with the turn, or after 45 seconds: resume it with the `Last-Event-ID` of the last event received. '400': description: >- Bad request: an invalid header, such as a reserved `X-End-User-Id` (`user_…`, `org_…`). content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: BAD_REQUEST message: >- X-End-User-Id must be an opaque, tenant-scoped id; ids with a reserved prefix, such as user_ or org_, are not allowed. '401': description: Authentication required content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: AUTH_REQUIRED message: Invalid or expired API key. doc_url: https://docs.gojinko.com/authentication/api-keys '403': description: >- The key is valid but is not a tenant API key (`jnk_t_…`), is not granted the `search_live` scope, or Travel AI search is not enabled for this account. **`FORBIDDEN`** — The credential is valid, and this request is refused anyway. Do not rotate the key — it is working. Check that it is entitled to this operation, and on this environment. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: FORBIDDEN message: >- This API key is not granted the "search_live" scope required by this endpoint '404': description: Resource not found content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: NOT_FOUND message: Resource not found. doc_url: https://docs.gojinko.com/concepts/errors '422': description: Validation error content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: BAD_REQUEST message: A required field is missing or invalid. doc_url: https://docs.gojinko.com/concepts/errors '429': description: Rate limit or quota exceeded headers: Retry-After: description: >- Seconds to wait before retrying. Absent when no interval applies. schema: type: string content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: RATE_LIMITED message: Too many requests. Retry after the Retry-After delay. doc_url: https://docs.gojinko.com/concepts/errors '502': description: The travel provider rejected the request, or an upstream call failed content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: UPSTREAM_REJECTED message: The request was declined for this offer. Try another offer. doc_url: https://docs.gojinko.com/concepts/errors '503': description: No travel provider can serve the request right now headers: Retry-After: description: >- Seconds to wait before retrying. Absent when no interval applies. schema: type: string content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: UPSTREAM_UNAVAILABLE message: >- The service is temporarily unavailable. Retry after the Retry-After delay. doc_url: https://docs.gojinko.com/concepts/errors '504': description: The travel provider did not answer in time content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' example: error: code: UPSTREAM_TIMEOUT message: >- The request timed out before a result was available. Retry the request. doc_url: https://docs.gojinko.com/concepts/errors security: - ApiKeyAuth: [] - BearerAuth: [] servers: - url: https://api.gojinko.com description: Production components: schemas: ErrorResponse: type: object properties: error: type: object properties: code: type: string enum: - INVALID_SEGMENTS_TO_EXCHANGE - PARTIAL_EXCHANGE_UNSUPPORTED - LEG_ALREADY_DEPARTED - DEPARTURE_DATE_IN_PAST - LEG_DATES_OUT_OF_ORDER - INVALID_EXCHANGE_REQUEST - NOT_EXCHANGEABLE - CREDENTIAL_FORMAT_INVALID - PAYMENT_TYPE_NOT_ENABLED - PAYMENT_CREDENTIAL_INVALID - TRIP_OWNED_BY_OTHER_PAYMENT - IDEMPOTENCY_KEY_REUSED - ATTEMPT_IN_PROGRESS - ATTEMPT_TERMINAL - TEMPORARILY_UNAVAILABLE - AUTH_REQUIRED - PAYMENT_REQUIRED - RATE_LIMITED - BAD_REQUEST - FORBIDDEN - NOT_FOUND - CONFLICT - GONE - QUOTE_PENDING - QUOTE_EXPIRED - TRIP_EXPIRED - TRIP_STATE_CONFLICT - OFFER_EXPIRED - OFFER_UNAVAILABLE - MISSING_CUSTOMER_DETAILS - INVALID_PHONE_NUMBER - INVALID_FLIGHT_NUMBER - CURRENCY_UNSUPPORTED - HOTEL_NAME_LOW_CONFIDENCE - DESTINATION_LOW_CONFIDENCE - UPSTREAM_REJECTED - UPSTREAM_UNAVAILABLE - UPSTREAM_TIMEOUT - UPSTREAM_ERROR - INTERNAL description: >- What went wrong, as a stable machine-readable code. This is a closed set — branch on it rather than on `message`, which is prose and may change. New codes arrive in a minor version, so treat an unknown one as its HTTP status. A code can also stop being emitted: it leaves this set in a minor version, named in the changelog, and a branch you wrote for it goes unreached rather than wrong. example: BAD_REQUEST message: type: string description: >- What went wrong, in prose. These codes always carry the same message, and it never includes text from a supplier: `RATE_LIMITED` — "Too many requests. Retry after the Retry-After delay."; `OFFER_EXPIRED` — "This offer has expired. Search again."; `OFFER_UNAVAILABLE` — "This offer is no longer available. Search again."; `UPSTREAM_REJECTED` — "The request was declined for this offer. Try another offer."; `UPSTREAM_UNAVAILABLE` — "The service is temporarily unavailable. Retry after the Retry-After delay."; `UPSTREAM_TIMEOUT` — "The request timed out before a result was available. Retry the request."; `UPSTREAM_ERROR` — "The request could not be completed.". doc_url: type: string field: type: string description: >- On a 400 `BAD_REQUEST` that names one refused input: the path of that field in your request, e.g. `selections[1].quantity` for the second selection of a `select_ancillaries` call. Absent when the refusal names no single field. example: selections[1].quantity required: - code - message required: - error securitySchemes: ApiKeyAuth: type: apiKey in: header name: X-API-Key BearerAuth: type: http scheme: bearer ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.