Image to Text (OCR) — Extract Text from Images and PDFs
Copy text out of photos, screenshots and scanned PDFs in 17 languages. Free OCR that runs in your browser — your images are never uploaded.
We're not interested in your data. A platform built free from surveillance.
Drop images or PDFs to extract their text
Choose filesJPG · PNG · WebP · HEIC · PDF
Works best on clear, straight photos, screenshots and scans. Handwriting is often not recognized.
Turn pictures of text into text you can edit
Photographed a page, a whiteboard or a sign? Took a screenshot of something you can't select? Drop it here and Localvert reads the words with OCR (optical character recognition) and gives you plain text to copy, search or paste into a document.
Choose the language of the text first. Recognition is much more accurate when the engine knows which alphabet and words to expect, and English is always included alongside the language you pick.
Advertisement
PDFs: instant when there's a text layer
Many PDFs already contain real text. In that case Localvert extracts it directly, which is instant and exact. Scanned PDFs are only pictures of pages, so each page is rendered and read with OCR — up to 30 pages at a time.
Private by design
OCR services usually upload your image to a server. Here the recognition engine (Tesseract, compiled to WebAssembly) runs inside your browser. The engine and language data are downloaded once from a public CDN and cached; your images and documents never leave your device — which matters for receipts, IDs, contracts and notes.
Getting the best results
Use good light and hold the camera straight above the page. Crop away anything that isn't text. Higher-resolution images help with small print. Printed text works far better than handwriting, and unusual fonts or text on busy backgrounds may need a cleaner photo.
Frequently asked questions
Which languages are supported?
English, Spanish, Portuguese, French, German, Italian, Japanese, Korean, Chinese (simplified and traditional), Russian, Vietnamese, Thai, Indonesian, Turkish, Arabic and Hindi.
Why is the first run slower?
The first time you use a language, the browser downloads the recognition engine and that language's data (a few megabytes). They are cached, so later runs start immediately.
Are my images uploaded?
No. Text recognition runs on your device. Only the engine and language files are downloaded; your files are never sent anywhere.
Can it read handwriting?
Neat block capitals sometimes work, but handwriting in general isn't reliable. The engine is designed for printed text.
Advertisement