Free OCR: Convert Scanned PDF to Editable Text, Word & TXT

Extract text from any scanned or image-based PDF and download it as a Word document or plain text file. Runs entirely in your browser — your documents never leave your device.

100% PrivateNo Sign-upFree Forever10+ Languages

Last updated: April 2026

What is OCR and what does this tool do?

OCR (Optical Character Recognition) is technology that converts images of text — like scanned documents or photos of printed pages — into editable, searchable text. This tool takes a scanned or image-based PDF, runs OCR on every page, and gives you back editable Word (.docx) and plain text (.txt) files.

Everything runs in your browser using WebAssembly (WASM). There is no server, no upload, and no sign-up. The Tesseract OCR engine is loaded once on first use (~2 MB) and then cached for instant re-use — the tool even works offline after that first load.

How it works

  1. 1

    Upload your scanned PDF

    Drag your PDF onto the drop zone or click to choose a file. Up to 25 MB and 50 pages per document.

  2. 2

    Choose the document language

    Pick one or more languages so Tesseract loads the right traineddata. Defaults to your browser language; supports 10+ languages.

  3. 3

    We extract the text in your browser

    Each page is rendered to a canvas and passed to Tesseract locally. You see progress, a live thumbnail, and the text as it streams in.

  4. 4

    Download as Word (.docx) or text (.txt)

    Review and inline-edit the extracted text, then download as Word, plain text, or both as a ZIP. Or send it straight into our translation service.

Supported languages

Run OCR in any of the languages below — including Spanish OCR, French OCR, German OCR, Chinese OCR, Japanese OCR, Korean OCR, Arabic OCR, and Russian OCR. Language data is fetched the first time you use a language and cached for future runs.

English
Spanish
Portuguese
French
German
Italian
Chinese (Simplified)
Japanese
Korean
Arabic
Russian

Is it really private?

Yes. This OCR tool runs 100% in your browser using WebAssembly. Your PDF is never uploaded to any server, we never see it, and no account is required.

If you then translate the extracted text using our document translation service, that upload is encrypted in transit and the original file is deleted within 7 days. OCR and translation are separate — OCR happens entirely on your device.

Common use cases

  • Digitizing scanned legal documents so they become searchable and editable.
  • Converting immigration paperwork — birth certificates, marriage certificates, diplomas — before submitting for translation.
  • Extracting text from historical records, old books, or typewritten archives for research.
  • Making scanned textbooks and lecture notes searchable for studying.
  • Preparing scanned business documents for translation into another language.

OCR accuracy tips

  • Use scans of at least 300 DPI. Low-resolution scans are the #1 cause of OCR errors.
  • Ensure pages are straight and not skewed. Even a small tilt hurts recognition accuracy.
  • Higher contrast between text and background improves results. Crank up contrast in your scanner settings if possible.
  • Select the correct language before processing. Running Spanish through an English-only OCR pass produces many errors.

Browser OCR vs. Cloud OCR vs. Desktop software

MethodPrivacyCostAccuracy
Browser OCR (this tool)100% private — runs locallyFreeGood on clean scans (≥ 300 DPI)
Cloud OCR (Adobe, Google)Files uploaded to third-party serversSubscription or pay-per-pageExcellent, but depends on vendor
Desktop softwarePrivate — runs on your machine$50–$500 one-time or subscriptionExcellent with handwriting add-ons

Already have editable text? Use our free Word Counter to count words and pages →

Need the extracted text translated?

We translate OCR'd documents into 30+ languages using DeepL neural machine translation. Click below and we'll pre-load your extracted text into the translation flow.

Translate a Document

Frequently asked questions

What is OCR?
OCR (Optical Character Recognition) is technology that converts images of text into editable, searchable text. If your PDF is a scan or a photo, its pages are just images — OCR looks at those images and figures out which characters and words appear on them so you can copy, edit, or translate the text.
Is this OCR tool really free?
Yes — no account, no credit card, no per-page charge. We offer OCR as a free companion to our paid document translation service, and most users never pay us a cent. If you eventually want to translate the extracted text, the translation step is what costs money.
Does my PDF get uploaded to a server?
No. Your PDF never leaves your browser. The OCR engine (Tesseract) runs entirely in your device using WebAssembly. You can verify this by opening the Network tab in your browser's developer tools while the tool runs — you'll see no requests to translateitnow.net during OCR.
How accurate is browser-based OCR?
Very accurate on clean, high-resolution printed scans (300 DPI or higher) — typically 95%+ word accuracy. Accuracy drops on low-resolution scans, skewed pages, very small fonts, unusual fonts, or poor contrast. We highlight low-confidence words in yellow so you can quickly spot and correct them before downloading.
What languages are supported?
Currently: English, Spanish, Portuguese, French, German, Italian, Chinese (Simplified), Japanese, Korean, Arabic, and Russian. You can select multiple languages for bilingual documents. Language models are downloaded the first time you use them and cached for future runs.
What's the maximum file size?
25 MB per PDF, and up to 50 pages. These limits are chosen to keep browser memory stable. For larger documents, split them into smaller PDFs using any free PDF splitter.
Can I OCR a handwritten document?
Limited. The Tesseract engine this tool uses is optimized for printed text — it struggles with cursive handwriting and produces lower accuracy even on block printing. For reliable handwritten OCR, a paid desktop tool with a dedicated handwriting model will do a better job.
Why did my OCR result have errors?
The three biggest causes are: (1) low scan resolution — anything under 300 DPI struggles, (2) skewed or rotated pages — even small tilts hurt accuracy, and (3) wrong language selected — running Spanish text through an English-only pass produces garbage. Re-scan at higher DPI, straighten the pages, and pick the right language.
Can I edit the text before downloading?
Yes. After OCR finishes, every page is inline-editable — click any page's text and fix errors directly. Your edits are used in the downloaded Word and text files and in the 'Translate this now' handoff.
How is this different from paid OCR like Adobe Acrobat?
Adobe Acrobat uses a cloud OCR engine that's slightly more accurate on difficult inputs (handwriting, low-quality scans, complex layouts), but it costs around $15/month and uploads your PDF to Adobe's servers. This tool is free, private (no uploads), and handles clean printed scans just as well — a good fit for everyday use.