For AI agents: a documentation index is available at the root level at /llms.txt and /llms-full.txt. Append /llms.txt to any URL for a page-level index, or .md for the markdown version of any page.
GuidesAPI ReferenceChangelogModel Versioning
GuidesAPI ReferenceChangelogModel Versioning
    • Getting Started
    • Authentication
    • API Versioning
    • SDKs
    • Deployments
    • Error Codes
    • Async Processing
  • Endpoints
      • POSTBatch Parse Files
      • POSTBatch Extract Files
      • POSTBatch Classify Files
      • POSTBatch Split Files
      • GETGet Batch Run
  • Webhook Events
LogoLogo
EndpointsBatch

Batch Classify Files

POST
/classify_runs/batch
POST
/classify_runs/batch
1import { ExtendClient } from "extend-ai";
2
3const client = new ExtendClient({ token: "YOUR_TOKEN" });
4await client.classifyRuns.createBatch({
5    classifier: {
6        id: "cl_xK9mLPqRtN3vS8wF5hB2cQ"
7    },
8    inputs: [{
9            file: {
10                url: "https://example.com/document1.pdf"
11            },
12            metadata: {
13                "customerId": "cust_abc123"
14            }
15        }, {
16            file: {
17                url: "https://example.com/document2.pdf"
18            },
19            metadata: {
20                "customerId": "cust_def456"
21            }
22        }, {
23            file: {
24                url: "https://example.com/document3.pdf"
25            },
26            metadata: {
27                "customerId": "cust_ghi789"
28            }
29        }]
30});
1{
2  "object": "batch_run",
3  "id": "bpr_Xj8mK2pL9nR4vT7qY5wZ",
4  "status": "PROCESSING",
5  "runCount": 50,
6  "createdAt": "2024-03-21T16:45:00Z"
7}
Submit up to **1,000 files** for classification in a single request. Each file is processed as an independent classify run using the same classifier and configuration. Unlike the single [Classify File (Async)](https://docs.extend.ai/2026-02-09/developers/api-reference/endpoints/classify/create-classify-run) endpoint, this batch endpoint accepts an `inputs` array and immediately returns a `BatchRun` object containing a batch `id` and a `PENDING` status. The individual runs are then queued and processed asynchronously. **Monitoring results:** - **Webhooks (recommended):** Subscribe to `batch_processor_run.processed` and `batch_processor_run.failed` events. The webhook payload indicates the batch has finished — fetch individual run results using `GET /classify_runs?batchId={id}`. - **Polling:** Call `GET /batch_runs/{id}` to check the overall batch status, and use `GET /classify_runs` filtered by `batchId` to retrieve individual run results. **Notes:** - A processor reference (`classifier.id`) is required — inline `config` is not supported for batch requests. - `inputs` must contain between 1 and 1,000 items. - All inputs in a batch use the same classifier version and override config.
Was this page helpful?
Previous

Batch Split Files

Next
Built with

Submit up to 1,000 files for classification in a single request. Each file is processed as an independent classify run using the same classifier and configuration.

Unlike the single Classify File (Async) endpoint, this batch endpoint accepts an inputs array and immediately returns a BatchRun object containing a batch id and a PENDING status. The individual runs are then queued and processed asynchronously.

Monitoring results:

  • Webhooks (recommended): Subscribe to batch_processor_run.processed and batch_processor_run.failed events. The webhook payload indicates the batch has finished — fetch individual run results using GET /classify_runs?batchId={id}.
  • Polling: Call GET /batch_runs/{id} to check the overall batch status, and use GET /classify_runs filtered by batchId to retrieve individual run results.

Notes:

  • A processor reference (classifier.id) is required — inline config is not supported for batch requests.
  • inputs must contain between 1 and 1,000 items.
  • All inputs in a batch use the same classifier version and override config.

Authentication

AuthorizationBearer

Bearer authentication of the form Bearer <token>, where token is your auth token.

Headers

x-extend-api-version"2026-02-09"Optional
API version to use for the request. If you're using an SDK, you can ignore this parameter. If you are not using an SDK and do not specify a version, you will either receive a `400 Bad Request` or be set to a previous legacy version. See [API Versioning](https://docs.extend.ai/2026-02-09/developers/api-versioning) for more details.

Request

This endpoint expects an object.
classifierobjectRequired
Reference to the classifier to run against every input in this batch.
inputslist of objectsRequired
An array of inputs to process. Each item produces one classify run. Must contain between 1 and 1,000 items.
priorityintegerOptional1-100Defaults to 50
An optional value used to determine the relative order of runs when rate limiting is in effect. Lower values will be prioritized before higher values.

Response

Successfully queued batch classify run
objectenum

The type of object. Always "batch_run".

Allowed values:
idstring

The unique identifier for this batch run.

Example: "bpr_Xj8mK2pL9nR4vT7qY5wZ"

statusenum

The status of a batch run:

  • "PENDING" - The batch has been created and is waiting to be processed
  • "PROCESSING" - The batch is currently being processed
  • "PROCESSED" - All runs in the batch have completed successfully
  • "FAILED" - The batch failed to process
  • "CANCELLED" - The batch was cancelled
Allowed values:
runCountinteger
The number of individual runs in this batch.
createdAtstringformat: "date-time"

The time (in UTC) at which the object was created. Will follow the RFC 3339 format.

Example: "2024-03-21T16:45:00Z"

Errors

400
Bad Request Error
401
Unauthorized Error
402
Payment Required Error
403
Forbidden Error
404
Not Found Error
422
Unprocessable Entity Error
429
Too Many Requests Error
500
Internal Server Error

API version to use for the request. If you’re using an SDK, you can ignore this parameter. If you are not using an SDK and do not specify a version, you will either receive a 400 Bad Request or be set to a previous legacy version. See API Versioning for more details.