SecondShiftAugie Discord Bot

AI-powered Discord bot with intelligent conversations and voice responses using VoxCPM TTS and Ollama AI.

Features

• AI Conversations: Intelligent responses using Ollama with Qwen2.5 model
• Voice Synthesis: High-quality text-to-speech using VoxCPM
• Slash Commands: Modern Discord command interface
• Voice Channel Integration: Automatic audio playback in voice channels
• Graceful Degradation: Continues working even when AI services are unavailable

Quick Start

1. Install Dependencies

   pip install -r requirements.txt
   

2. Configure Environment

   cp .env.template .env
   # Edit .env with your Discord bot token and channel IDs
   

3. Prepare Reference Files

   • Place your reference audio file at assets/model.wav
   • Place the corresponding transcript at assets/transcript.txt

4. Start the Bot

   python run_bot.py
   

Commands

The bot uses modern Discord slash commands:

• /join - Join your current voice channel
• /play - Replay the last generated voice response
• /leave - Leave the current voice channel
• /help - Show available commands and usage information
• /health - Show detailed bot and system health status

Usage

1. Join a voice channel in Discord
2. Use /join to have the bot join your channel
3. Mention the bot (@SecondShiftAugie) in chat
4. The bot will respond with both text and AI-generated voice

Configuration

Required Environment Variables

• BOT_TOKEN: Your Discord bot token
• CHANNEL_ID: Discord channel ID where the bot will send messages

Optional Environment Variables

• VOICE_CHANNEL_ID: Default voice channel ID for the bot to join
• SAVE_PATH: Directory for temporary audio files (default: ./temp_audio)

VoxCPM Configuration

• VOXCPM_MODEL_PATH: Model path or Hugging Face model ID (default: openbmb/VoxCPM-0.5B)
• VOXCPM_PROMPT_WAV: Path to reference audio file (default: assets/model.wav)
• VOXCPM_PROMPT_TEXT: Path to reference text file (default: assets/transcript.txt)

Ollama AI Configuration

• OLLAMA_BASE_URL: Ollama API endpoint (default: http://localhost:11434)
• OLLAMA_MODEL: AI model to use (default: qwen2.5:1.7b)
• SYSTEM_PROMPT_FILE: System prompt configuration file (default: ollama_system_prompt.json)