OCR Text Extractor
Pull text out of photos, screenshots and scanned PDFs in English, Hindi, Bengali, Tamil and other Indian languages. Text recognition runs in your browser with Tesseract, so your files stay on your device.
- Runs on your device
- No sign-up
- Free, no watermark
Guide: How to Extract Text From Images and PDFs Using Online OCR
How to use the OCR Text Extractor
- 01Choose the languages in your document, up to 3. English is chosen at first, and your choice is remembered.
- 02Drop images or PDF files into the box, or click it to choose them. iPhone HEIC photos are converted to JPG on your device first.
- 03Press Extract text. The first time an image or scanned page needs OCR, the engine and your languages download before reading starts.
- 04Watch the progress of each file. PDFs are read page by page. Press Cancel to stop.
- 05Read the extracted text, then copy it or download it as a .txt file.
Good to know
Your files stay on your device
Images and PDFs are read in your browser. They are never uploaded.
Downloaded once
The OCR engine and each language you choose (1 to 6 MB each) come from the jsDelivr CDN the first time, then your browser caches them.
Images and PDFs
Read JPG, PNG, WebP, iPhone HEIC and other images, or every page of a PDF, one after another.
15 languages
English, Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Odia, Urdu, Assamese, Nepali and Sanskrit. Up to 3 at once.
Frequently asked questions
Are my images or PDFs uploaded?
No. Your files are read and recognised in your browser and never leave your device. The only download is the OCR engine itself: the Tesseract program and the data for each language you choose are fetched once from the jsDelivr CDN and then cached by your browser. No file content is sent with those requests.
Why is the first run slower?
The first time OCR is needed, your browser downloads the OCR engine and the data for each language you chose, between 1 and 6 MB per language. Later runs use the cached copy and start much faster; adding a new language downloads only that language. PDFs that already contain text don't need the engine at all.
Which languages are supported?
English, Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Odia, Urdu, Assamese, Nepali and Sanskrit. Choose up to 3 at once, for example Hindi and English for a form with both. Each language is a separate download and makes reading slower, so choose only the ones in your document. Printed text works best; handwriting is usually not recognised well. If you change the languages after reading a file, press Extract text to read it again.
Can it read scanned PDFs?
Yes. Each page is rendered as an image and read in turn, and the text of all pages is put together in order. If a page already contains selectable text, that text is copied directly instead, which is faster and exact; only pages without text go through OCR. The result says which was used. Password-protected PDFs must be unlocked first.
How can I get better results?
Use sharp, well-lit images with good contrast and horizontal text. Crop away busy backgrounds, and make sure small text is large enough to read. Handwriting is usually not recognised well.