Skip to main content

Re-Extract Document

Re-run the extraction pipeline on an existing document without re-uploading the file. Use this to apply a new schema or updated instructions to a previously processed document.

POST/v1/documents/:id/re-extract

Parameters

ParameterTypeRequiredDescription
idstring (path)YesThe document ID to re-extract
schemastring (JSON)NoNew inline schema for extraction
schema_idstringNoNew schema ID reference
instructionsstringNoUpdated extraction instructions

Example

curl -X POST https://api.talonic.com/v1/documents/doc_xyz789/re-extract \
  -H "Authorization: Bearer tlnc_sk_live_7f3a...x9k2" \
  -H "Content-Type: application/json" \
  -d '{"schema_id": "sch_abc123"}'

Behavior

Re-extraction creates a new extraction and job. The original extraction is preserved. The four-phase pipeline runs again with the new schema, producing fresh per-cell provenance. Use this after updating a schema or adjusting the confidence gate. See schema formats for format options.

Related

Extract Endpoint
Primary extraction entry point.
Extractions API
Query extraction results.
Jobs API
Track extraction job progress.
Schema Formats
Inline JSON, schema_id, and auto-detect.
Four-Phase Pipeline
How documents are processed.