Real-time voice to voice Gemini discord chat bot

Chat voice to voice live with Google's Gemini 2.0 AI with google search capabilities.

Demo

scarlett.mp4

Prerequisites

• Python 3.8 or higher
• Discord Bot Token
• Free GEMINI API Key: https://aistudio.google.com/
• Discord Server with voice channels enabled

Installation

1. Clone the repository:

git clone https://github.com/2187Nick/discord-voice-to-voice-gemini
cd discord-voice-to-voice-gemini

2. Install required packages:

pip install -r requirements.txt

3. Create a .env file in the project root with the following variables:

DISCORD_TOKEN=your_discord_bot_token
GEMINI_API_KEY=your_gemini_api_key

4. Options:

• Set the voice to use in main.py line 22.

voice="aoede"

• Set the persona to use in main.py line 23.

persona="Take on the persona of an overly excited motivational speaker"

Usage

1. Start the bot:

python main.py

2. In Discord, join a voice channel and use the following command:

/chat

3. Enable push-to-talk in Discord and hold the key while speaking.

4. Interrupt the response by pressing the key again and start speaking.

Commands

• /chat - Initiates a voice chat session with the bot
• /stop - Stops the current voice chat session

Support

https://x.com/2187Nick

http://discord.gg/vxKepZ6XNC

Project Structure

├── main.py # Bot initialization and command handling
└── src/ 
    ├── record.py # Audio processing and speech-to-text conversion 
    ├── stream.py # Custom audio streaming implementation 
    └── gemini.py # Gemini AI WebSocket client integration

Technical Details

• Uses Discord.py for bot functionality
• Implements custom audio streaming for real-time voice processing
• Uses WebSocket connection to Gemini AI for real-time responses
• Handles both synchronous and asynchronous operations for optimal performance