Skip to main content
Submit a list of document URLs to Atlas for asynchronous bulk OCR extraction. Atlas queues each document, processes it, and stores the results against the returned batch_id. You then use that batch_id with GET /bulk/extracts to retrieve the extracted data once processing is complete.
The Bulk Service does not support callbacks. You must poll GET /bulk/extracts to retrieve results.

Endpoint

Authentication

This endpoint requires a JWT token obtained from POST /bulk/authtoken. Pass it in the Token request header.

Request headers

string
required
JWT access token issued by POST /bulk/authtoken.

Request body

string
required
Product type identifier for this batch. Example: "POC_CDL".
array
required
List of document file objects to process. Each element must include both file_url and document_id.

Response

A 201 Created response contains an array — one entry per uploaded document — each sharing the same batch_id.
string
The batch identifier assigned to all documents in this upload (e.g., "BATCH-30732-SIYKJ-20251121113228"). Use this value when polling GET /bulk/extracts.
string
Echo of the document_id you provided for this file.
string
Internal UUID Atlas assigned to the individual file. Useful for support queries.

Error responses

Examples

Save the batch_id immediately — all documents in the same upload share it, and you need it to fetch results.