yuta1984

yuta1984 / ndlocrlite-web

Public

NDLOCRLite Web: ブラウザ完結型日本語OCRツール（ONNX Web Runtime使用）

43
2
100% credibility
Found Mar 01, 2026 at 43 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

A browser application that extracts Japanese text from images and PDFs using on-device processing with model caching for privacy.

How It Works

1
🌐 Discover the tool

You find a free online Japanese text reader that works right in your web browser, perfect for pulling words from photos or scanned pages.

2
📁 Drop in your files

Drag and drop your image files, PDFs, or even paste pictures from your clipboard – it handles single photos or whole folders easily.

3
✨ Start the reading magic

Click the big button to begin, and it previews your pages so you can check before it dives in.

4
⏳ Watch it work

A friendly progress bar shows it finding text areas and reading the words, grabbing smart pieces the first time for speed later.

5
👀 See highlighted results

Your images appear with colored boxes around detected text, click any area to zoom and read exactly what it found.

6
📋 Review and reuse

Browse full text, past sessions from history, or tweak settings if needed, all saved safely on your device.

🎉 Copy your text

Grab the clean Japanese text to paste anywhere or download it as a file, mission accomplished without sending anything online.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 43 to 43 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is ndlocrlite-web?

ndlocrlite-web is a TypeScript web app that runs Japanese OCR entirely in your browser using ONNX Runtime, processing images, PDFs, or folders without sending data to servers. Drag-drop files, paste from clipboard, or select regions for targeted recognition—it extracts text blocks, reorders them logically, and outputs copyable full text with previews. Solves offline digitization of Japanese docs for privacy-focused users.

Why is it gaining traction?

Zero server dependency means instant privacy and offline use after initial model cache, unlike cloud OCR APIs with costs and quotas. Handles multi-page PDFs natively, bilingual UI (Japanese/English), history storage, and region re-OCR for fine-tuning—devs love the drag-select workflow over clunky desktop tools. TypeScript polish and Vite speed make it a snappy prototype base.

Who should use this?

Japanese historians scanning old manuscripts, researchers extracting text from NDL scans, or web devs prototyping doc-analysis UIs. Ideal for frontend teams needing quick Japanese text extraction in SPAs without backend hassle.

Verdict

Grab it for local Japanese OCR experiments—features deliver real value despite 42 stars and 1.0% credibility signaling early maturity. Fork and contribute; docs are README-basic but code runs smooth out-of-box.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.