ScanLex · document recognition and OCR

Convert PDF to Word

Turn PDFs, scans and photos into editable Word, Excel and plain text — tables and layout included. Our desktop app works offline.

For companies with a heavy document flow there is ScanLex Server: the same recognition engine deployed on your own infrastructure, with no user limit.

14 days · no sign-up · Windows and Linux
INPUT
PDF
PDF14 pages
PDF, JPG, TIFF
or a photo of a document
OUTPUT
W Word · DOCXtext, layout, footnotes
Excel · XLSXtables as real cells
TXT Text · TXTplain text
PDF+ Searchable PDFsame look, now searchable

One scan — any of these formats on the way out. Everything is computed on your own computer.

The same output formats, processed on your own server — files never leave your perimeter.

Choose how you want to work

The desktop app and ScanLex Server are built on the same recognition engine — the same accuracy and layout, at a different scale.

PC

Desktop app

offline · 1 computer

Recognition runs entirely on your PC — files are never sent anywhere. The right choice for sensitive documents and high-volume scanning.

  • Fully offline — documents never leave your computer
  • Suitable for contracts, invoices and ID documents
  • No limits on file size or number of files
  • Built-in viewer, saves to Word / Excel / TXT / PDF
  • 14 days free, then a 1- or 2-year licence for one computer

Windows 10 / 11, Astra Linux, ALT Linux · free trial, no payment or sign-up required.

S

ScanLex Server

on-premise · the whole organization

The same recognition engine — only deployed on your own infrastructure. Data never leaves the company perimeter, and there is no cap on the number of users.

  • Data never leaves your perimeter — fit for personal data (Russian Federal Law 152-FZ)
  • No user limit — you pay by server capacity
  • Browser access — your staff install nothing
  • Russian-made software — a domestic replacement for enterprise OCR
  • Deployed in a single working day

Pricing goes by server core count, with no limit on users.

Accurate exactly where OCR usually breaks

Text, tables, fonts and layout are carried over just like the original.

scan
Excel
table: scan → Excel

Tables stay tables

Borders, merged cells and numeric columns come through as structure rather than as a picture — you can recalculate the totals straight away.

contract layout

Layout and columns

Headings, lists, footnotes and multi-column text keep their order and styles.

editing right on the scan

A PDF editor on top of the scan

Need to change a single line — edit it right on the image: stamps, signatures and paper stay the very same pixels, and the result is saved as PDF.

The formats you need

Word, Excel, plain text and searchable PDF with a text layer.

Russian and English

Mixed-language documents — with no manual mode switching.

Photos of documents

Auto-rotation, skew correction and upscaling of small type before recognition.

Check it before you save

The result is shown right in the app — page through it and only then save.

Three steps to a finished file

Open it, recognize it, save it. No complicated settings.

1

Add a file

PDF, JPG, PNG, TIFF, BMP, WEBP — multi-page scans and phone photos. Drag the file into the app window.

2

Check the result

The recognized document is displayed right in the app — page through it before saving.

3

Save it in your format

Word, Excel, TXT or PDF with a text layer. Tables and layout are preserved.

For business

Documents never leave your perimeter

Neither the app nor the server sends files to any external service. That is exactly why ScanLex is chosen where online converters are forbidden by internal policy.

Request a pilot
152-FZFit for documents that contain personal data
0 bytesLeave for the internet during recognition
RussiaDeveloped in Russia — a domestic replacement for enterprise OCR
1 dayTo deploy the server and train your staff

Who uses this every day

Lawyers

Contracts and annexes from PDF into Word — so you edit the wording instead of retyping it.

Accounting

Invoices, waybills and acts: tables go into Excel as real cells and the totals recalculate straight away.

Archives

Digitizing paper files: multi-page scans and searchable PDFs with a text layer.

Public sector

Handling personal data where online converters are forbidden by internal policy.

Pricing and licences

Free trial

$0
14 days · no sign-up

The full app — check the accuracy on your own documents.

Download
MOST POPULAR

Licence for 1 computer

$16.36 / year
one computer · updates included
$27.69 / 2 years
one computer · better value
$37.75 / year
one computer · paid by invoice
$62.92 / 2 years
one computer · paid by invoice

No limits on the number or size of files, with updates for the whole licence term.

The same features, paid by invoice for your organization. Leave a request — we will send the invoice and payment details.

Buy a licence Buy a licence Request an invoice Request an invoice

ScanLex Server

from $629
per year · by core count · no user limit

A pilot on your own infrastructure and a quote for your document flow.

Request a quote

PDF to Word and OCR — frequently asked questions

ScanLex is optical character recognition (OCR) and document conversion: PDF to Word, scan to Word, photo to text, PDF to text, table recognition into Excel. The desktop app is for home and work; ScanLex Server is for organizations with a heavy document flow.

How do I convert a PDF to Word?

Install the ScanLex app on your computer (there are builds for Windows, Astra Linux and ALT Linux) and open the file — it recognizes the text and tables and saves the result as an editable Word (DOCX) document with the layout preserved. It works offline and is free for 14 days.

How accurately are tables recognized?

Tables come through as structure rather than as a picture: borders, merged cells and numeric columns become a real Word or Excel table, so you can recalculate the totals straight away. Tables with ruled lines work best — invoices, waybills and standard accounting forms. In borderless forms the structure is reconstructed from how the text is placed, so it is worth reviewing the result before you save it.

Can it handle multi-page PDFs?

Yes. The desktop app has no limit on the number of pages or the file size — scans hundreds of pages long are handled, and for a large document you can recognize just the range you need. In the online service the page volume depends on your plan.

Does it work without an internet connection?

Yes. The desktop app recognizes documents right on your computer and needs no permanent connection: the internet is only required to download the app and to activate the licence once. ScanLex Server also runs in an isolated network, with offline licence activation.

Which languages are supported?

Russian and English, including both in the same document — there is no mode to switch by hand. Contracts with Latin brand names, part numbers and e-mail addresses are recognized in full.

How do I recognize text from a scan or a photo?

ScanLex reads text from photos and scans (PDF, JPG, PNG, TIFF, BMP, WEBP) and moves it into Word, Excel or TXT with tables and formatting intact. It handles photos taken with a phone camera too.

What happens to my documents after recognition?

In the desktop app nothing is sent anywhere: both the original and the result stay on your computer and we never see them. In the online service uploaded files are stored temporarily and deleted automatically after 7 days. With ScanLex Server documents never leave the organization perimeter.

Can I use the app on more than one computer?

A licence covers one computer, so several workstations need several licences — you can buy them in a single order. The 14-day trial is also granted per computer. When there are many workstations, ScanLex Server is usually the better deal: it is installed on one server and staff work through a browser with no cap on the number of users.

How does the desktop app differ from Server?

The recognition engine is exactly the same — what differs is how you run it. The app is installed on a computer (Windows, Astra Linux or ALT Linux), works offline and is licensed per workstation. ScanLex Server is deployed on your own server, staff open it in a browser and install nothing, there is no cap on users, and the licence is counted by the number of server cores.

Try it on your own documents

14 days of the full app — with no payment and no sign-up. For organizations there is a ScanLex Server demo on your own infrastructure.

We use cookies and Yandex.Metrica web analytics so that the site works properly and we can see how it is used. By continuing to use the site, you agree to this and to our Privacy Policy.