What is Tecack?
Tecack is a handwriting recognition engine built with TypeScript.
This project was developed based on KanjiCanvas.
The character candidate inference algorithm was implemented by KanjiCanvas,
and this project focuses on setting up the surrounding environment.
✅ TypeScript support
✅ Installable via npm
✅ Package segmentation
✅ Comprehensive documentation
✅ Rich ancillary tools (such as dataset creation tools)
and more...
The Excellence of KanjiCanvas
👍 Retention of stroke order data
👍 Inference based on stroke order and number of strokes
👍 Extensibility of datasets
👍 Lightweight
How does it work?
Tecack is broadly divided into frontend and backend.
The frontend generates TecackStrokes based on user's stroke information through Canvas. The backend infers character candidates based on TecackStrokes.
The backend can run in any environment that supports JS/TS. This includes browsers, as well as Node.js, Deno, and Bun. (You can think of it as a pure function that takes TecackStrokes and a dataset, and outputs character candidates.)
