Complex documents. Clear, verifiable data.
Extract the fields you need from PDFs, scans, Office files, and spreadsheets. Keep the page, quote, and verification status with every value.
Playground
Parse a document or extract fields with page-level evidence.
Request
Fixed example input for this walkthrough.
As code
curl -X POST https://factget.dev/docs/v1/extract \
-H "Authorization: Bearer $KEY" \
-H "Content-Type: application/json" \
-d '{
"url": "https://example.com/invoice.pdf",
"fields": [
{
"name": "invoice_number",
"type": "string"
},
{
"name": "issue_date",
"type": "date"
},
{
"name": "total",
"type": "number"
}
]
}'Response
Sample3 fields found invoice.pdf · Page 1
invoice_numberINV-2026-1042Found
“Invoice INV-2026-1042”Page 1 · Supporting quote
issue_date2026-10-30Found
“Issued: 30 October 2026”Page 1 · Supporting quote
total1280Found
“Total due: USD 1,280.00”Page 1 · Supporting quote
Interactive console preview with sample data. No live jobs or monitors are created.
Read the document. Keep the evidence.
Define the fields in words or with a JSON Schema. Docs parses the file, checks the extraction, and flags the fields that need a closer look.
Explore the API referenceText layers are read directly. Scanned pages are transcribed and checked against the page image.
Use a field list, a JSON Schema, or a plain-language description of what you need.
Found, not found, ambiguous, invalid, and missing statuses keep doubtful fields visible.
A few things to know.
How do I provide a document?
Upload a file through the API or send a public document URL. Uploaded files can be referenced by fileId in parse and extract requests.
How large can a file be?
Docs supports files up to 50 MB. Uploaded files are retained for 24 hours for parsing and extraction.
Can I extract only certain pages?
Yes. Use the pages parameter with a range such as 1-10 or a selection such as 3,7,12-15.
Put Docs to work for your team.
Bring us the research you do by hand. We’ll show you what FactGet can take off your plate.