PDFConverterOnline.com

Free · runs in your browser · nothing is uploaded · works offline

OCR a Scanned PDF Online — Searchable PDF, Text or Word, No Upload

Recognise the text in a scan on your device and get a searchable PDF, plain text or Word.

Recognise the text in a scan on your device and get a searchable PDF, plain text or Word.

      Private by design: processing happens on your device with open-source libraries served from this site (licences). Nothing is uploaded, and once loaded the tool keeps working offline. Need other tools? All tools are on the homepage.

      Free, no ads, no upsell. If it saved you a trip to a paid converter, you can buy the developer a coffee — it funds new tools and keeps this one free. ☕ Buy me a coffee

      Runs on your deviceTested on Pixel 7 & MacBook Air M2Open-source engines · licences

      About this tool

      Scanned PDFs are pictures of pages, so you cannot search or copy their text. This tool runs the open-source Tesseract engine inside your browser (WebAssembly) and gives back a searchable PDF — your original pages with an invisible text layer — and optionally a .txt or Word file. Eleven languages are available, each alone or paired with English; Indian and Arabic scripts are embedded with proper fonts so they are searchable too.

      The engine (≈4 MB) and each language pack (1–2 MB) download once from this site and are then cached for offline use. Clean 300 DPI scans reach about 95% accuracy; expect 1–4 seconds per page.

      How it works

      1. Step 1. Choose or drop scanned PDFs or photos of documents.
      2. Step 2. Pick the language, the output (searchable PDF, text, Word or all) and 200 or 300 DPI.
      3. Step 3. Press Run OCR. The note shows words found and average confidence.

      Questions

      Is this really offline OCR?

      Yes. After the first run the engine and language data are stored in your browser; you can disconnect and keep recognising text. Nothing is ever sent to a server.

      Does OCR change my original pages?

      No. The searchable PDF keeps every page image exactly as it was and adds an invisible text layer on top.

      Which language should I pick for a bilingual document?

      Choose the "+ English" variant of the other language (for example Hindi + English).

      Good to know

      • Handwriting and very low-resolution photos score poorly.
      • Pages must be upright; use Scan cleanup first for tilted scans.

      Updated · Part of PDFConverterOnline.com, the independent guide to every PDF converter online.