Telegram Voice Transcription Bot

A Telegram bot that automatically transcribes voice messages into text using OpenAI's Whisper model.

Features

• Automatic transcription of voice messages to text
• Support for both groups and private conversations
• Automatic activation when added to a group
• Automatic cleanup of temporary files
• Audio quality optimization before transcription

Prerequisites

• Python 3.11+
• FFmpeg
• A Telegram bot token (obtained from @BotFather)

Installation

Using Docker (recommended)

1. Clone the repository:

git clone https://github.com/[your-username]/telegram-voice-transcription-bot.git
cd telegram-voice-transcription-bot

2. Configure your Telegram token in config.py or via environment variable

3. Build and run the container:

docker build -t voice-transcription-bot .
docker run -d voice-transcription-bot

Manual Installation

1. Clone the repository:

git clone https://github.com/[your-username]/telegram-voice-transcription-bot.git
cd telegram-voice-transcription-bot

2. Install dependencies:

pip install -r requirements/dev.txt
# Or for production
pip install -r requirements/prod.txt

3. Install FFmpeg:

• On Debian/Ubuntu:

  sudo apt-get install ffmpeg

• On macOS:

  brew install ffmpeg

4. Configure your environment:

   • Create a .env file in the project root directory
   • Add your Telegram bot token:

     TELEGRAM_BOT_TOKEN=your-token-here
     
     LOGGING_LEVEL=INFO # Logging level (DEBUG, INFO, WARNING, ERROR)
     WHISPER_MODEL=large-v3 # Model name (large-v3, base, etc.)
     WHISPER_DEVICE=cpu # Device (cpu, cuda)
     WHISPER_COMPUTE_TYPE=int8 # Compute type (int8, fp16, fp32)
     
     AUDIO_LANGUAGE=fr # Language code (en, fr, es, etc.)
     AUDIO_INITIAL_PROMPT=Transcription of a French voice message. # Initial prompt for Whisper

5. Run the bot:

python main.py

Usage

1. Start a private conversation with the bot or add it to a group

2. In private conversation:

   • Send a voice message
   • The bot will automatically transcribe the message

3. In group:

   • Add the bot to the group
   • The bot activates automatically
   • Send a voice message
   • The bot will reply with the transcription

Technologies Used

• python-telegram-bot - Framework for Telegram bots
• faster-whisper - Optimized implementation of Whisper
• pydub - Audio manipulation