Turn any document into structured data
PDF to JSON
Read a PDF page by page and get clean JSON in the shape you define — text-based exports and scanned PDFs alike.
What goes in
Single or multi-page PDFs, whether generated by software or scanned from paper.
- Every page read, with page numbers kept against each value.
- Tables become arrays of line items, not one block of text.
- Each field links back to where it was found, so you can check it.
What comes out
{
"source": {
"file_name": "supplier-statement-sep.pdf",
"pages": 3,
"type": "pdf"
},
"document_type": "statement",
"account": {
"supplier": "Vermeer Materialen BV",
"account_number": "NL-44192"
},
"period": {
"from": "2026-08-01",
"to": "2026-08-31"
},
"currency": "EUR",
"entries": [
{
"page": 1,
"date": "2026-08-04",
"reference": "INV-90211",
"amount": 1840.5
},
{
"page": 2,
"date": "2026-08-18",
"reference": "CRN-00412",
"amount": -212
}
],
"closing_balance": 1628.5
}How it works
01
Upload a file or take a photo. PDFs, photos and scans all work, up to 100 MB on paid plans.
02
Formalini reads it and fills your template, or proposes a structure from the document itself.
03
Check the values against the page, then export the JSON or send it onwards by API or webhook.
Common questions
Does it work on scanned PDFs with no text layer?
Yes. Scanned pages are read as images, so a PDF made from a photocopier works the same way.
How large can the file be?
Up to 100 MB on paid plans. The free no-sign-up demo takes the first three pages of a file up to 10 MB.
How to convert a PDF to JSON in four steps
01
Upload your PDF
Drop a text-based or scanned PDF into the free demo — no account needed for the first three pages.
02
Pick a structure or let it detect one
Choose a template such as invoice or statement, or let Formalini detect the fields and tables automatically.
03
Check the values
Every value is linked to its page and spot on the page, with a confidence score, so you only review what is uncertain.
04
Export or send the JSON
Download the JSON or CSV, or send it to your system by webhook or API on every new document.
PDF to JSON questions
- How do I convert a PDF to JSON for free?
- Use the free demo on this site for up to three pages without signing up, or create a free account for 100 pages a month.
- Can I use an LLM to extract data from a PDF?
- Yes — Formalini uses AI reading models, but adds what a plain chatbot lacks: a fixed output structure, validation, confidence scores and a link from every value back to the page.
- Does it keep tables and line items?
- Yes. Tables become JSON arrays of rows, with each row keeping its page number.
- Can I automate PDF to JSON with an API?
- Yes. Send files to the upload API and receive structured JSON back, or get it pushed to your webhook.