Jobs & batch

Submit a batch

POST /api/v1/extract/batch

Multipart form: multiple files, plus the same extract options as sync extract (extract_tier, custom_fields, system_prompt, …).

Constraints:

• Lite is not supported for batch yet (use Plus/Pro/Ultra).
• Max 100 total pages across all files in the batch.
• Credits are charged up front for the sum of pages.

Response:

{
  "batch_id": "...",
  "status": "processing",
  "jobs": [{ "job_id": "...", "status": "pending", "filename": "..." }],
  "total_pages": 12,
  "message": "Batch submitted. Poll /batches/{batch_id} for status."
}

Poll a job

GET /api/v1/jobs/{job_id}

Statuses: pending | processing | completed | failed.

On completion, results may include extracted_json, extract_metadata, citations, filename (payload depends on retention settings).

Poll a batch

GET /api/v1/batches/{batch_id}

Returns the batch status and nested job statuses.

History

GET /api/v1/history

Lists recent extracts for the authenticated org/user (shape aligned with upload responses).

Delete a document

DELETE /api/v1/documents/{document_id}

Requires auth with org or user scope.