Real-time Audio Transcription

Purpose

This project aims to provide a real-time audio transcription system, where audio input from a microphone is sent to a server, transcribed, and then displayed to the user in real-time. The project uses a server to do the heavy calculations to do the actual transcription while a lightweight client just does the audio recording and another client just does the result display.

Python Files

transcribe_server.py

This file contains the server-side logic, which:

• Listens for incoming audio chunks from the client
• Transcribes the audio chunks using whisper
• Returns the transcribed text to the client

audio_grabber.py

This file contains the client-side logic, which:

• Captures audio from the microphone
• Chunks the audio into manageable pieces
• Sends the audio chunks to the server with a unique chunk ID

HTML Files

transcribe_listener.html

This file contains the client-side logic, which:

• Listens to the server for transcribed chunks
• Displays the transcribed text to the user in real-time

Setup and Run

To set up and run the project, follow these steps:

• Install the required Python packages: pyaudio, flask, requests, whisper
• Run audio_grabber.py to start capturing audio from the microphone
• Run transcribe_server.py to start the server
• Open transcribe_listener.py in the browser to start displaying transcribed text in real-time

./server -m models/ggml-large-v3.bin -l de -p 16 -t 32 --host 0.0.0.0 --port 8007