PDF editor for Windows
Scanned books, made readable, searchable and easy to move through.
BixoPDF fixes faint scans, reads the text in many languages on your own PC, and turns a book's contents page into bookmarks and working links. It leaves out the rest, so it opens fast and stays out of your way.
Windows 10 (1809+) and 11 · 64-bit · No account, no subscription
Marine Insurance Act 1906
1. A contract of marine insurance is a contract whereby the insurer undertakes to indemnify the assured, in manner and to the extent thereby agreed, against marine losses, that is to say, the losses incident to marine adventure.
2. A contract of marine insurance may, by its express terms, or by usage of trade, be extended so as to protect the assured against losses on inland waters or on any land risk which may be incidental to any sea voyage.
Small, light and fast
Big PDF suites carry years of features most people never open. We kept what you do with documents every day and removed the rest, so the app starts quickly and pages turn without waiting, even in a file of thousands of pages.
A real document
A court expert report · 3,286 pages · 383 MB
- On screen, ready to scroll
- under 1 s
- Jump to page 3,286
- 0.2 s
- Ready to edit
- seconds later, in the background
Measured on our own Windows 11 PC, opening the file from disk with a click.
Kept
- Read, search and copy text, even on scanned pages
- Highlight, draw, write, stamp and sign
- Merge, split, crop, rotate and reorder pages
- OCR that runs on your own computer
- Bookmarks and contents links built automatically
- Compress, password-protect, save pages as images
Left out
- Sign-in, cloud storage and sync
- Monthly subscription and renewal reminders
- Upsell banners and trial nags
- Form designers and print-production preflight
- Menus nested five levels deep
- Your documents leaving your PC
Fix a faint scan in seconds
Old photocopies and grey scans are tiring to read. Move a slider and the page changes on screen at once. When it looks right, apply it to one page, a range like 1-5, or the whole book.
Thickness
Thin, broken strokes become solid letters again.
Sharpness
Soft, out-of-focus edges become crisp.
Contrast
Grey text on grey paper becomes black on white.
Brightness
Yellowed or dark pages become clean paper.
Before you commit. The preview shows the page as it will be. Nothing is changed until you apply it.
Page ranges. Type 1,2,3 or 10-40 to apply the same settings to many pages at once.
Also on the page. Crop margins, split two-page spreads into single pages, and straighten tilted scans for OCR.
OCR that we trained further ourselves
We started from open-source recognizers and kept training them on our own data. On a scanned book in Korean BixoPDF made about a third of the errors of leading PDF program A and a fifth of program B's. On a book in English it matched the better of the two. It all runs on your PC, so your documents never go to a server.
36 scanned pages of a Korean court handbook
53,123 characters · body text, boxed examples, tables and 6-pt footnotes
| Program | Characters wrong | Bar |
|---|---|---|
| BixoPDF | 0.84% | |
| Leading PDF program A | 2.31% | |
| Leading PDF program B | 4.20% |
Bars on a 0–5% scale. Lower is better.
32 scanned pages of an English construction law book
81,443 characters · body text, quotations and footnotes
| Program | Characters wrong | Bar |
|---|---|---|
| BixoPDF | 0.25% | |
| Leading PDF program A | 0.26% | |
| Leading PDF program B | 0.35% |
Bars on a 0–0.5% scale. Lower is better.
How we tested. Pages of born-digital PDFs were turned into 300-dpi greyscale scans with slight tilt, blur, noise and JPEG compression, saved as image-only PDFs, and given to each program's OCR on the same PC in October 2026. Each result's text layer was read the same way and compared with the source text, spaces left out. Neither book was used to train our recognizers. Programs A and B are two widely used commercial PDF editors with built-in OCR, at the versions installed on the test PC.
Disclaimer. These results come from tests run in a limited environment on a limited set of documents. They do not mean that the same results will be obtained in every environment or with every document. Accuracy depends on the scan, the layout, the fonts, the language and the settings used, and your results may differ.
Latin script · default
English, French, German, Spanish, Portuguese, Italian, Dutch, Danish, Norwegian, Swedish, Finnish, Icelandic, Polish, Czech, Slovak, Hungarian, Slovenian, Croatian, Serbian (Latin), Bosnian, Lithuanian, Estonian, Turkish, Albanian, Catalan, Irish, Indonesian, Malay, Tagalog, Swahili and Latin
Korean
Hangul, Hanja and word spacing in one pass, with Japanese quotations.
Chinese
Simplified and traditional, on a current-generation recognizer.
Japanese
Kanji and kana, with Chinese and English on the same page.
Arabic
Right-to-left lines put back in reading order.
Searchable PDFsAn invisible text layer under the page image, so search and copy work in any PDF reader.
Batch OCRDrop a folder of PDFs and come back to all of them searchable.
Straightens firstTilted pages are deskewed before reading, for cleaner lines.
Replaces weak OCRText an older program hid under the scan is swapped for the new result.
35 OCR languages · all included
- Western Europe
- English, French, German, Spanish, Portuguese, Italian, Dutch, Catalan, Irish
- Northern Europe
- Danish, Norwegian, Swedish, Finnish, Icelandic
- Central and Eastern Europe
- Polish, Czech, Slovak, Hungarian, Slovenian, Croatian, Serbian (Latin), Bosnian, Albanian
- Baltic
- Lithuanian, Estonian
- Other Latin-script languages
- Turkish, Indonesian, Malay, Tagalog, Swahili, Latin
- East Asia
- Korean (Hangul and Hanja), Chinese (simplified and traditional), Japanese
- Right to left
- Arabic
Latin-script languages are listed when every letter they use is in the recognizer's alphabet. Romanian, Latvian, Vietnamese, Maltese, Esperanto, Welsh and Azerbaijani use some letters it does not have, so those letters may be read as similar ones.
Bookmarks and contents links, made for you
A 1,000-page scan is only useful if you can get around it. BixoPDF reads the book's own contents pages and builds the bookmark tree and the links for you. You check it and correct the few lines it is unsure of.
Point to the contents pages
Pick them in the page list, for example pages 5–12.
It reads every entry
Title, level from the numbering, and printed page number. Printed pages are matched to PDF pages through an offset or the book's page labels, then checked against each page's text.
Correct with a few clicks
Doubtful lines are flagged. Change a level with + and −, fix a page number, include or skip a line.
Make bookmarks and links
Each contents line becomes a link to its chapter. The links work in other PDF readers too, and running it again replaces the old links instead of stacking new ones.
And the everyday tools
Annotate
Highlight, underline, shapes, text boxes, memos, stamps and signatures.
Pages
Insert, delete, move, rotate, crop and split pages; set page labels like i, ii, 1, 2.
Merge
Combine many PDFs and images into one file, from the app or Explorer's right-click menu.
Compress
Shrink scans with black-and-white, grayscale, JPEG quality and resolution, with a size estimate first.
Search many files
Find a word across a whole folder of PDFs and open the page it is on.
Protect and export
Password-protect with AES, save pages as JPG or PNG, export OCR text.
Make big scans small
Scanned books run to hundreds of megabytes, too big to e-mail, upload to a court or client portal, or keep by the thousand. Choose how far to go, see the size and the page before you commit, and save a smaller copy.
A real book
The Law of Marine Insurance · 1,082 scanned pages
45% smaller, with its colour pages kept in colour. Text pages turned to black and white can shrink to a tenth or less.
Nothing else is lost. Text, OCR text, annotations and bookmarks stay as they were. Only the page images are recompressed.
Your original stays. The result is saved as a new file and opens in its own tab, so you can compare.
Search still works. A compressed OCR'd book is still searchable and copyable in any PDF reader.
One price: $29
No tiers and no monthly bill. Every feature on this page is included.
BixoPDF · full version · Microsoft Store
- Scan enhancement, OCR, bookmarks and links
- All OCR languages included
- Runs offline, no account needed
- Windows 10 (1809+) and 11, 64-bit
- Updates through Microsoft Store
Most PDF suites are sold by subscription, and a year of one costs several times this price.
BixoPDF is paid for once, and does the jobs that matter most for scanned documents: making them readable, searchable and easy to move through.
Questions
Short answers to what people ask first.
Does OCR need an internet connection?
No. The recognizers are installed with the app and run on your computer. Your documents are not uploaded anywhere.
Which Windows versions are supported?
Windows 10 version 1809 or later, and Windows 11, on 64-bit PCs.
Will my bookmarks and links work in other PDF readers?
Yes. They are saved as standard PDF bookmarks and link annotations.
What happens to text an older OCR program left in my PDF?
If that text is hidden under the page image, BixoPDF replaces it with the new OCR text when you save. If the text is the visible content of the page, it asks you first.
Can I make the app open my PDFs by default?
Yes. In Windows Settings → Apps → Default apps, choose BixoPDF for .pdf files.
Where do I buy it, and how do updates arrive?
BixoPDF is sold on Microsoft Store. It is installed and kept up to date by the Store, on every Windows PC signed in with the same Microsoft account.
How do I get help?
Write to support@bixopdf.com. The support page lists what to include.