This project provides a real-time audio transcription system, sending microphone audio to a server for processing and displaying the transcribed text to users. It's designed for applications requiring live speech-to-text capabilities, benefiting users who need immediate textual representation of spoken words.

How It Works

The system employs a client-server architecture. A client (audio_grabber.py) captures microphone audio, segments it into chunks, and transmits these chunks to a server (transcribe_server.py). The server utilizes the Whisper model for transcription and returns the results to a separate client (transcribe_listener.html), which displays the text in real-time. This separation offloads heavy computation to the server, keeping clients lightweight.

Quick Start & Requirements

• Install dependencies: pip install pyaudio flask requests
• Run server: python transcribe_server.py
• Run audio grabber: python audio_grabber.py
• Open listener: transcribe_listener.html
• Server command example: ./server -m models/ggml-large-v3.bin -l de -p 16 -t 32 --host 0.0.0.0 --port 8007
• Requires Python 3.x, PyAudio, Flask, Requests, and Whisper.

Highlighted Details

• Real-time audio transcription via client-server architecture.
• Utilizes Whisper for transcription.
• Separate clients for audio capture and result display.

Maintenance & Community

No information on contributors, community channels, or roadmap is available in the README.

Licensing & Compatibility

The README does not specify a license. Compatibility for commercial or closed-source use is undetermined.

Limitations & Caveats

The project is presented as a basic implementation without explicit error handling, advanced features, or performance benchmarks. The README lacks details on supported operating systems, hardware requirements beyond standard Python environments, or specific Whisper model compatibility.