Get Document Fields
Return a document's captured fields bound into the field registry — one row per occurrence with its canonical concept, tier, value, and confidence.
Returns the fields captured from a document and bound into the field registry. Where the extraction endpoints give you the raw field-name/value pairs a document produced, this endpoint gives you the registry-resolved view: each captured value is reported alongside the canonical concept it bound to, that concept's tier, and the confidence of the binding. It is the structured result most integrations actually want per document.
Each row carries the occurrence value and confidence, the resolved registry field (its id, canonical name, display name, and data type), the registry tier (1 = canonical, 2 = provisional, 3 = candidate), and the semantic cluster the field belongs to when one is assigned. Rows are ordered by tier, so the most established concepts come first.
/v1/documents/{id}/fieldsResponse
Response fields
curl
Response
{
"document_id": "7f3a1b2c-0000-0000-0000-000000000000",
"data": [
{
"field_id": "b2c3d4e5-0000-0000-0000-000000000000",
"canonical_name": "invoice_total",
"display_name": "Invoice Total",
"cluster_name": "amounts",
"data_type": "number",
"tier": 1,
"value": "490.00",
"confidence": 0.97
}
]
}