English日本語
KotenOCR

KotenOCR

古典籍・くずし字を、スマホで読む。 Read classical Japanese texts on your phone.

国立国会図書館のAIモデルを活用した、完全オフラインの古典籍OCRアプリ。 Fully offline OCR app powered by the National Diet Library's AI models.

📷

撮って認識 Snap & Recognize

カメラで古典籍を撮影するだけ。くずし字を自動で検出・認識します。 Just photograph a classical text. Kuzushiji characters are automatically detected and recognized.

🔒

完全オフライン Fully Offline

すべての処理はデバイス上で完結。画像やデータが外部に送信されることはありません。 All processing happens on-device. No images or data are ever sent externally.

高速処理 Fast Processing

ONNX Runtimeによる高速推論。撮影から数秒で認識結果を表示します。 Fast inference via ONNX Runtime. Results appear within seconds of capture.

📋

テキストコピー Copy Text

認識結果はワンタップでコピー。研究や学習にすぐ活用できます。 Copy results with one tap. Ready for your research and studies.

使い方 How It Works

1

撮影 Capture

カメラまたはフォトライブラリから画像を選択 Take a photo or select from library

2

認識 Recognize

AIが文字の位置を検出し、くずし字を翻字 AI detects characters and transcribes kuzushiji

3

活用 Use

結果をコピー・エクスポートして研究に活用 Copy or export results for your research

利用技術 Technology

NDL古典籍OCR-Lite

CC-BY-4.0 — 国立国会図書館

GitHub

ONNX Runtime

MIT License — Microsoft

GitHub