Fill a schema from a document by URL
Fill an arbitrary user JSON schema from a document URL, with citations.
Authorizations
Body
Input to a schema-extraction request.
Given a document and an arbitrary user JSON schema, fill the schema's fields from the document and cite where each value came from.
HTTP(S) URL of the document to extract from. PHI-enabled keys must use the file-upload route instead.
JSON Schema describing the fields to extract; field descriptions are instructions the extractor follows. Required unless auto_schema is true.
What happens to a value whose citation cannot be verified against the document. true (default): the value is nulled out and its path listed in ungrounded_fields, so a fabricated value never reaches you. false: the value is kept but still flagged in ungrounded_fields.
Set true (and omit schema) to have a schema designed from the document first, then filled with the same grounded extraction. The schema used is returned in generated_schema.
Include figures from the parse stage in the extraction context. Set false to extract from text and tables only.
false (default): response unchanged. true: additionally return ocr_text — the whole parsed document as a single text string, concatenated in reading order (the same text POST /v1/parse returns as content).
Response
Successful Response
What this request charged, in credits (plan 078 D6).
Present only on responses whose request was actually charged — absent
(never null) on unbilled lanes (demo, legacy-PHI ledger) so pre-credits
response shapes stay byte-identical. credits = pages × credits_per_page
at the v1 card: parse 1.0, schema extract 4.0 all-in.