Extract Text from Photos and Scans — Free OCR for English + Hindi
Image-to-text conversion in your browser, no upload. Works on scanned IDs, exam papers, receipts, handwritten notes (sort of), and supports both English and Hindi.
You have a photo of an exam syllabus, a scanned bank statement, or a handwritten note. You want to copy the text out — but typing 200 lines by hand is not happening.
That's what OCR (Optical Character Recognition) is for. And it now runs in your browser.
What our OCR tool does
Upload one or more images (JPG / PNG / WEBP). The tool runs them through Tesseract.js — a WebAssembly port of Google's Tesseract OCR engine — entirely on your device. No upload.
Output: clean text you can copy or download as .txt.
Supported languages: - English - Hindi (हिन्दी) - English + Hindi combined (for documents that mix both)
What works well
| Input | OCR quality | |---|---| | Printed PDF text photographed clearly | ★★★★★ Excellent | | Scanned book pages | ★★★★☆ Very good | | Newspaper articles | ★★★★☆ Very good | | Receipts (clear) | ★★★☆☆ Good | | Whiteboard notes | ★★★☆☆ Good | | Neat handwriting | ★★☆☆☆ Some errors | | Cursive / messy handwriting | ★☆☆☆☆ Don't try |
When to use it
Copy text from a scanned PDF
If you only have an image of a PDF (no embedded text), our PDF text extractor won't find anything. OCR is the answer:
1. Convert PDF pages to images using PDF to JPG. 2. Run each image through OCR. 3. Combine the resulting .txt files.
Capture exam syllabus from a photo
Took a photo of the notice board listing your exam syllabus? OCR it, paste into your notes.
Digitize old documents
Decades-old certificates, marksheets, photos of newspaper clippings — OCR them once, search them forever.
Translate scanned text
OCR → copy text → paste into Google Translate. (We don't translate inside the tool because that would require a server call.)
First-run download warning
The first time you use the OCR tool, it downloads ~5-15 MB of language data per language. After that, it's cached and runs instantly. The English model is the smallest; English + Hindi combined is the largest.
Tips for best results
1. Crop tightly to the text — fewer non-text pixels means less for the OCR engine to misread. 2. Increase contrast if the original is faded — use the Photo Enhancer first. 3. Convert color photos to B&W for cleaner recognition (the Document Scanner does this automatically). 4. Straighten skewed scans — Tesseract handles small angles, but rotate manually if it's more than ~5°.
For the cleanest workflow on phone scans: scan with the Document Scanner (which auto-applies B&W + contrast), then OCR the output.