SnapDoc
Scanned Arabic PDFs, turned into perfect editable Word
6 followers
Scanned Arabic PDFs, turned into perfect editable Word
6 followers
SnapDoc is an Arabic-first OCR tool that turns scanned or photographed Arabic PDFs and images into fully editable Word files. Most converters break connected letters, flip the reading order, and turn tables into a mess. SnapDoc preserves right to left layout, connected letters, tashkeel, tables, and colors, so government forms, legal contracts, and invoices come out clean and editable. Bilingual Arabic and English supported. Free at 1 page a day.
Free
Launch Team / Built With
Render: Get up to $100K in startup creditsThe fastest way to deploy and scale apps and agents
Promoted
Hi Hunters,
I am the founder of SnapDoc. I built it after watching every OCR tool I tried do unspeakable things to Arabic: letters divorced from each other, whole lines reading backwards, and tables that came out looking like abstract art. Charming in a gallery, less charming on a contract, a research paper, or a twenty year old archived file.
Arabic is genuinely hard for machines. It reads right to left, letters change shape depending on their mood and position, tashkeel floats above the words, and tables love to flip themselves for fun. Most tools run one OCR pass, shrug, and hand you the wreckage. SnapDoc is Arabic-first and actually does the work:
- Reconnects letters and preserves shaping, so words stay words, not ransom notes
- Keeps tashkeel attached where it belongs, not sprinkled around like confetti
- Rebuilds right to left flow and RTL tables, so the first column sits on the right where an Arabic reader expects it and the total lands in the correct cell, not three rows away
- Repairs broken encodings and quietly deletes the characters the OCR hallucinated, instead of passing the nonsense on to you
- Carries over colors, bold, and layout, and runs confidence gates, so when it is unsure it gives you plain correct text instead of a confident wrong guess
You get a real editable Word file, not a screenshot cosplaying as a document. Bilingual Arabic and English pages too, which is basically all paperwork in the Gulf.
It is built for anyone whose work is trapped inside scanned Arabic: lawyers buried in contracts, researchers digitizing old texts, writers and translators reusing their own material, government offices, and companies sitting on archives of paper nobody wants to retype.
I will not claim perfection, no OCR is flawless, and anyone who says otherwise is selling something. But the rule here is correct text over a pretty guess, which is exactly what serious documents need. It is live and free at 1 page a day.
So please, send me your worst. The blurriest phone photo, the crookedest scan, the invoice that has been through a printer three times. The messier the better, I want to see where it breaks.