Extract data from
fixed-layout documents.

Draw a box around any region of a PDF or scanned image and DocPeel pulls out exactly the text inside — no templates, no schemas, no AI guesswork. Built for invoices, lab reports, ID cards, shipping labels, and every form your team works with on repeat.

Start for freeSee it in action
  • No credit card required
  • Runs in your browser
  • Works on PDFs & images
What is Zonal OCR?

The simplest way to grab a single value from a fixed-layout document.

Zonal OCR — short for zone-based optical character recognition— lets you outline a rectangular region of a document and pull out only the text inside that box. It's the perfect complement to AI extraction when you need a deterministic, pixel-precise answer for documents that always look the same: invoices from the same vendor, identity cards, lab forms, shipping labels, even fax-scanned PDFs.

  • Pixel-precise. The OCR engine reads only what you box, never the surrounding text.
  • Schema-friendly. Each zone gets a field name, producing clean JSON output ready for your pipeline.
  • Browser-native. No backend roundtrip — OCR runs locally with WebAssembly so your data stays put.
field_a
field_b
How it works

From document to clean data in three steps.

No setup, no template configuration, no AI training. Just point, click, and extract.

1

Upload your document

Drop a PDF or scanned image into DocPeel. Multi-page PDFs are supported, and you can paginate through them inside the Zonal OCR canvas.

+ pdf
2

Draw a zone over the data

Click and drag to mark the region containing the value you need. Zones snap to the document at full resolution for pixel-accurate OCR results.

INVOICE #INV-0481TOTAL$3,402.00invoice_no
3

Submit fields to your extraction

Name each zone — e.g. invoice_no, total — and click Submit. Fields are merged into your extraction's JSON, ready to send to Sheets, Airtable, or your API.

{"invoice_no":"INV-0481""total":"$3,402.00"}
When to use Zonal OCR
AI
Zonal
field

Pair it with AI extraction for the best of both worlds.

DocPeel's flagship AI extraction handles whole documents end-to-end and is unmatched for variable-layout content. Zonal OCR is the surgical scalpel for fixed-layout cases — and the easiest way to audit, correct, or supplement what the AI returned.

AI extraction
  • Variable layouts & new vendors
  • Whole-document understanding
  • Tables, line items, multi-page
Zonal OCR
  • Identical layouts every time
  • One specific value or total
  • Audit / correct existing fields
Use cases

Built for documents your team sees every day.

Recurring invoices

Pull the invoice number, date, and total from vendor invoices that always use the same layout.

Receipts & expense reports

Read totals and merchant names from scanned receipts uploaded by your team.

ID cards & passports

Extract document numbers and expiry dates from KYC document scans without a custom model.

Shipping labels

Capture tracking IDs, carrier codes, and weights from labels printed in standard formats.

Lab & medical reports

Read specific result fields from forms with consistent layouts — no PHI ever leaves the browser.

Government forms

Tax forms, customs declarations, and licenses where the same boxes appear in the same place every time.

FAQ

Frequently asked questions

Regular OCR reads the entire document and returns a wall of text. Zonal OCR limits the recognition to a rectangle you draw, so the output is exactly the field you asked for — no parsing, no regex, no cleanup needed.

Try Zonal OCR on your next document.

Upload a PDF or image, draw a box, get clean JSON. No template setup, no credit card.

Start for free