WhisperLive Client

A lightweight client for real-time transcription and translation with the WhisperLive server.

Features

• Real-time audio transcription/translation.
• Supports Whisper models and Voice Activity Detection (VAD).
• Streams audio to the server; optionally saves .wav output.

Setup

1. Install Python 3.8+ and dependencies. Make sure portaudio or its development package e.g. portaudio19-dev is installed.

   • via pip:

   pip install loguru numpy pyaudio websocket-client

2. Start the WhisperLive server.

Usage

Run the client:

python client.py

Configure options (server, task, output) in the __main__ block.