Image OCR (Text Recognition)
Extract text from photos and screenshots in your browser.
Fully browser-based — no server upload, no registration.
Drag & drop an image
or click to select
JPEG, PNG, WebP, BMP supported (max 30MB)
📋 Paste from clipboard with Ctrl+V
About
The "Image OCR (Text Recognition)" tool is a free online tool that reads text from photos and screenshots in your browser and extracts it as editable text.
It uses Tesseract.js (WebAssembly version), which runs Google's industry-leading Tesseract OCR engine directly in the browser. It supports 5 languages including Japanese, English, Simplified Chinese, Korean, and Spanish.
All processing runs in a Web Worker inside your browser — no image data is ever sent to a server. Use preprocessing features (contrast enhancement, grayscale, binarization) to improve accuracy on low-quality images.
How to Use
Upload an Image
Drag and drop, click to select, or paste from clipboard (Ctrl+V). Select the target language checkboxes before processing.
Adjust Preprocessing
In the "Preprocessing" tab, adjust contrast, binarization threshold, and grayscale to improve accuracy. Then click "Extract Text".
Copy or Save Results
The extracted text appears in the "Result" tab. Edit if needed, then copy to clipboard or download as a UTF-8 TXT file.
Glossary
- OCR (Optical Character Recognition)
- Technology that automatically recognizes text in images or scanned documents and converts it to editable digital text. Modern OCR engines leverage deep learning for high accuracy.
- Tesseract.js
- A JavaScript library that runs Google's open-source Tesseract OCR engine in the browser via WebAssembly. Supports 100+ languages. No server required.
- WebAssembly (WASM)
- A binary instruction format that allows running native code (C/C++/Rust) at near-native speed in the browser. Enables computationally intensive tasks like OCR to run client-side.
- Binarization
- Converting each pixel to either black or white based on a threshold value. Applying binarization to images with low contrast between text and background can significantly improve OCR accuracy.
- Confidence Score
- A 0–100% metric indicating how confident the OCR engine is in its recognition result. Higher values indicate more reliable results. Manual review is recommended when below 70%.
FAQ
- Q.Are my images uploaded to a server?
- No. Tesseract.js (WebAssembly) runs entirely inside your browser. No image data is transmitted. Only the language model is downloaded from a CDN on first use.
- Q.What is the accuracy for Japanese text?
- For clear printed text (typeset), expect 85–95% accuracy. Preprocessing (contrast, binarization) can improve this. Handwritten characters have significantly lower accuracy.
- Q.Why is the first load slow?
- The language model (approx. 15MB for Japanese) is downloaded from a CDN on first use. After that, it's cached in the browser — subsequent uses load instantly.
- Q.Can it recognize handwriting?
- Tesseract.js is optimized for printed text. Handwriting recognition accuracy is limited. For high-accuracy handwriting OCR, consider Google Docs or specialized services.
- Q.Can I extract text directly from a PDF?
- PDFs are not supported. Take a screenshot or export the PDF page as a PNG/JPEG image first, then use this tool.
Use Cases
Digitizing Documents
Extract text from scanned contracts or reports to create searchable digital documents.
Signs & Menus While Traveling
Extract text from foreign-language signs or menus to paste into a translation app.
Screenshot Text Extraction
Extract text from non-copyable PDFs, videos, error message screenshots, or locked documents.
Textbook & Study Notes
Extract text from photos of textbooks or study materials to paste into digital notes.
Business Card Digitization
Extract names, companies, and contact info from business card photos to add to your address book.
Research & Archiving
Extract text from images of old documents or research materials to build a digital archive.