Enqueue a document for AI extraction (human-review required)

Stores the uploaded source document tenant-scoped, runs deterministic + AI extraction, and enqueues the resulting transaction CANDIDATES in a human-review queue. AI output is NEVER authoritative: no transaction is reconciled until a human approves the review. The tenant is resolved from the JWT and the context/source from the path — never from the body.

POST /v1/imports/contexts/{contextId}/sources/{sourceId}/extract-document

Enqueue a document for AI extraction (human-review required)

cURL

curl --request POST \
  --url https://matcher.sandbox.lerian.net/v1/imports/contexts/{contextId}/sources/{sourceId}/extract-document \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/pdf' \
  --data '<string>'

202

Example

{
  "candidateCount": 12,
  "reviewId": "550e8400-e29b-41d4-a716-446655440000",
  "status": "PENDING_REVIEW"
}

Autorizaciones

Authorization · string · header · requerido

Bearer token authentication (format: "Bearer {token}")

Parámetros de ruta

contextId · string<uuid> · requerido

Context ID

sourceId · string<uuid> · requerido

Source ID

Cuerpo

application/pdf

The body is of type file.

Respuesta

202

application/json

Accepted

candidateCount · integer<int64> · requerido

Number of transaction candidates queued for review

Ejemplo:

12

reviewId · string<uuid> · requerido

Identifier of the queued extraction review

Ejemplo:

"550e8400-e29b-41d4-a716-446655440000"

status · enum<string> · requerido

Lifecycle status of the queued review (always PENDING_REVIEW on enqueue)

Opciones disponibles: PENDING_REVIEW

Ejemplo:

"PENDING_REVIEW"