Classify

Classify labels a document by type, with the text that supports the label. Classify is a server feature.

curl -s -X POST http://localhost:8000/classify -H "content-type: application/json" -d '{
  "input": "reducto://doc.pdf",
  "classification_schema": [
    {"category": "Invoice", "criteria": ["has an invoice number", "has a total due"]},
    {"category": "Receipt", "criteria": ["shows a payment already made"]},
    {"category": "Contract", "criteria": ["has parties and signatures"]}
  ],
  "model": "accurate"
}'

With Reducto's SDK: client.classify.run(input=..., classification_schema=[...]).

Field	Meaning
classification_schema[]	category and its criteria
category_groups	optional grouping of categories
model	default, or accurate: the quoted evidence is checked against the page text in the same call
page_range	classify from some pages only
document_metadata	extra context, for example where the file came from