Free pilot Book a demo
Product · Document extraction

Document extraction

Capture every field and table row from any business document, with a confidence score on each value.

Book a demo → Start a free pilot

What it handles

  • Scans, phone photos and PDFs
  • Typed and printed text in many layouts
  • Header fields, tables and line items
  • New formats without building templates

What you get

  • Structured JSON or CSV for every document
  • A confidence score on each field
  • The source line highlighted for quick checks
  • Accuracy that improves from reviewer corrections
What is included

Built for real-world documents

Template-free reading

Add a new supplier or form type without configuring anything. The model reads the layout in front of it.

Tables and line items

Multi-page tables, merged cells and wrapped descriptions come back as clean rows.

Confidence on every value

Set thresholds so only uncertain fields reach a person.

See the output

What comes back, and what gets checked

Each document returns structured fields with confidence scores, plus the results of your validation rules. Anything uncertain goes to review.

Illustrative example with sample data.

VendorNorthwind Traders
Invoice no.INV-20418
Total₹ 1,84,200.00
Line items12 rows captured
All required fields foundPass
Totals match line itemsPass
Date format unusualReview
How it works

From document to data in four steps

1

Collect

Documents arrive by email, upload, SFTP or API.

2

Read

Sigixtract classifies, splits and extracts the fields and tables.

3

Check

Your rules and cross-checks run on every document.

4

Deliver

Clean data goes to your systems. Exceptions go to review.

FAQ

Common questions

Do I need to train the model first?

No. You can start on your own documents on day one. Reviewer corrections then help the model improve over time.

Which file types can I send?

PDFs, scans and common image formats. Email, upload, SFTP and API intake are all supported.

Related

Product

Classification & splitting

Send one big file. Sigixtract works out what is inside it, separates each document and routes it to the right workflow.

 Product

Validation & review

Check extracted data against your rules, then give people a fast screen to resolve only what is uncertain.

 Document

Invoice data extraction

Extract vendor details, dates, taxes, totals and every line item from invoices in any layout.

See Sigixtract on your own documents

Send us a real invoice, statement or form. We will show you every field extracted and checked.

Book a demo → Explore products