WebSocket

In this section, we describe how to use the WebSocket server and client for real-time speech recognition with sherpa-onnx.

The WebSocket server is implemented in C++ with the help of websocketpp and asio.

Hint

It does not depend on boost.

It does not depend on boost.

It does not depend on boost.

• Streaming WebSocket server and client
  • Build sherpa-onnx with WebSocket support
  • View the server usage
  • Start the server
  • View the usage of the client (C++)
  • Start the client (C++)
  • View the usage of the client (Python)
  • Start the client (Python)
  • Start the client (Python, with microphone)
• Non-streaming WebSocket server and client
  • Build sherpa-onnx with WebSocket support
  • View the server usage
  • Start the server
    • Start the server with a transducer model
    • Start the server with a paraformer model
  • Start the client (Python)
    • offline-websocket-client-decode-files-paralell.py
    • offline-websocket-client-decode-files-sequential.py