Voice Note Transcription and Summarizer Bot for Telegram 🎙️📝

A powerful Telegram bot that transcribes and summarizes voice notes using state-of-the-art AI models. Built with Python and powered by Groq's API with Whisper and Llama 3 model for transcription and summarization.

Screenshots 📸

Transcription Example	Summary Example

Features ✨

• Transcribe voice notes from forwarded messages
• Handle direct voice note recordings
• Generate accurate transcriptions using Whisper
• Provide concise summaries of the transcribed content using Llama 3
• Support for multiple audio formats
• Well-formatted, easy-to-read output

Prerequisites 📋

• Python 3.10 or higher
• pipenv (Python package manager)
• Telegram Bot Token (from @BotFather)
• Groq API Key (Get it here)

Installation 🚀

1. Clone the repository:

   git clone https://github.com/aviaryan/voice-transcribe-summarize-telegram-bot.git
   cd voice-transcribe-summarize-telegram-bot
   

2. Install dependencies using pipenv:

   pipenv install
   

3. Create a .env file in the root directory:

   cp .env.copy .env
   

4. Fill in your environment variables in the .env file:

   TELEGRAM_BOT_TOKEN=your_telegram_bot_token
   GROQ_API_KEY=your_groq_api_key
   

5. Configure authorized users:

   • Open bot.py and locate the AUTHORIZED_USERS array
   • Add your Telegram user ID to the array (you can get your ID by messaging @userinfobot on Telegram)

   AUTHORIZED_USERS = [your_telegram_id]  # Add more user IDs as needed
   

Usage 🎯

1. Activate the virtual environment:

   pipenv shell
   

2. Start the bot:

   python bot.py
   

3. In Telegram:

   • Forward any message containing a voice note to the bot
   • Record and send a voice note directly to the bot
   • Wait for the bot to process and return both transcription and summary

How it Works 🔄

1. The bot receives a voice note through Telegram
2. Audio is processed and sent to Groq's API
3. Whisper model transcribes the audio content
4. Another pass through Groq's API generates a concise summary using Llama 3 model
5. Both transcription and summary are returned to the user in a well-formatted message

Contributing 🤝

Contributions are welcome! Feel free to:

• Open issues for bugs or feature requests
• Submit pull requests
• Improve documentation
• Share feedback

License 📄

This project is licensed under the MIT License - see the LICENSE file for details.