日本語 | English

This project is a Node.js server that uses VOSK to recognize speech and send it over WebSocket.

1. Clone this repository.
2. Navigate to the project root directory.
3. Run npm install to install dependencies.
4. Download and extract VOSK model files and put them in the model directory. The directory structure should look like the following:

[sskw]$ ls -l model
合計 24
-rw-r--r-- 1 sskw sskw  898  7月  9  2022 README
drwxr-xr-x 2 sskw sskw 4096  7月  9  2022 am
drwxr-xr-x 2 sskw sskw 4096  7月  9  2022 conf
drwxr-xr-x 3 sskw sskw 4096  7月  9  2022 graph
drwxr-xr-x 2 sskw sskw 4096  7月  9  2022 ivector
drwxr-xr-x 2 sskw sskw 4096  7月  9  2022 rescore

1. Navigate to the project root directory.
2. Run npm start to start the server.

Once the server is started, it runs a WebSocket server on port 8080. When a client connects to the WebSocket, recognized text is sent to the client.

To pause the sending of recognition results, press the space key. To resume, press the space key again.

MIT License