InoTools Dev

Image Text Extractor (OCR)

A free tool to extract text from images with high accuracy using AI OCR. Supports multiple languages for scanned documents and screenshots. Works entirely in your browser, ensuring your images are never uploaded to any external server.

Drop image here or click to select

Supports JPEG, PNG. You can also paste an image from your clipboard (Ctrl+V / Cmd+V).

Guide & Terminology

What is the Image Text Extractor (OCR)?

Optical Character Recognition (OCR) uses AI to detect and extract written text from images, photos, and scanned documents so it can be edited or copied digitally.

How to Use

  1. Drop an image or paste it from your clipboard using Ctrl+V / Cmd+V.
  2. Select the optimal engine (General Purpose or Japanese Specialized) based on the image's source material.
  3. If the image features light text on a dark background, enable "Invert colors" to improve recognition accuracy.
  4. Click "Extract Text" to run the local AI model.

Terminology

  • NDLOCR-Lite: An open-source model optimized by the Japanese National Diet Library, exceptionally good at reading complex Kanji and vertical text.
  • Tesseract: A highly versatile open-source engine capable of reading standard characters in over 100 languages.

Limitations / Notes

  • The first time you run this tool, it will securely download the required AI models (approx. 80MB) directly into your browser's persistent cache.
  • Highly stylized fonts or extremely blurry images may produce hallucinated text.

Security

All OCR models are executed locally via WebAssembly inside your sandbox browser. Your images are never uploaded to any server.