Document to Data

Any document into Excel tables, clean text and JSON

Every heading, paragraph, table, key-value pair, checkbox and signature in reading order, each shown on its page, ready for spreadsheets and AI.

Start free Contact sales

Works on

Bank statement · page 1 · table 1

Date	Description	Debit	Balance
02/10	Transfer PT Mitra		125,400,000
03/10	Payroll	48,200,000	77,200,000
05/10	PLN	3,150,000	74,050,000

How it works

4 steps, on the same files

01
Upload

One PDF of up to 50 pages, or 1 to 10 photos, text or scan.
02
Read the document

Optionally split a combined PDF into its documents.
03
Check it on the page

Blocks are drawn over the page; poor scans and unverified blocks are marked.
04
Download

Excel with one sheet per table, CSV, Markdown or JSON, or go on to Extract Data.

What you get

Used in

Real tables

Cells with their spans, tables that continue across pages joined, numbers kept as numbers and IDs as text.
Correct a cell

Fix a value and the Markdown and downloads follow.
Scans too

Pages are straightened and read with OCR before the layout is worked out.
Mostly free

Text and OCR use no credits; AI checks only tables without lines whose rows are unclear, 1 credit a page.
Ready for AI search

Markdown and chunks for RAG, from the API as well.
Safe spreadsheets

CSV that cannot run as a formula when opened.

Start free Contact sales