Documents

Read the matter's documents into its facts

Requires the documents scope. Queues an extraction job and answers 202 at once; poll GET /jobs/{id} or listen for extract.completed. Each document is read in full, several are merged, and the result is written to the matter's case_context (where the app keeps confirmed facts) with a per-document record_index. Without document_ids, every document the key can see that is included in analysis is read. Free. Document text is never stored, only the structured facts. Pages that are pictures of text are not read through the API, so a scanned document fails as unreadable. With a test key the job finishes at once with a fixed sample and no model is called.

post/matters/{id}/extract

Request

  • Base URL: https://api.suprememind.ai/v1
  • URL: https://api.suprememind.ai/v1/matters/{id}/extract
  • Auth: HTTP bearer

Headers

Idempotency-Keystring

Makes a POST safe to retry for 24 hours.

Request body

document_idsinteger[]
domain_hint'securities' | 'antitrust'

Response

The extraction was queued (or, for a test key, finished).

idstring required
object'job' required
kind'extract' | 'report' required
livemodeboolean
status'queued' | 'succeeded' required
matter_idstring
created_atstring date-time required

Changes

No recorded changes to this operation across all 1 revision of this API.