🦁 Beast Mode

⚠️ Alpha / Proof of Concept — This project is in early alpha and is not production-ready. It represents a working concept and exploration of ideas around AI-assisted development environments. Expect rough edges, missing features, and breaking changes. Contributions, feedback, and experimentation are welcome — but treat this as a starting point, not a finished product.

An AI-powered coding agent with voice-to-text capabilities, built on VS Code's open-source foundation. Beast Mode allows you to control your development environment through voice commands and leverage powerful AI models via Ollama for autonomous coding tasks.

✨ Features

• 🎤 Voice-to-Text Integration: Control your IDE with voice commands using Whisper
• 🤖 AI-Powered Coding Agent: Autonomous code generation, refactoring, and debugging
• 🔄 Multiple AI Models: Switch between different Ollama models on the fly
• ⚡ Autonomous Execution: Execute commands without intervention based on prompts
• 🎯 Wake Word Detection: Activate with "beast" or customize your own wake word
• 📝 Full File System Control: Create, edit, and delete files through voice or text
• 🔧 Command Execution: Run terminal commands directly from voice prompts
• 🌐 Open Source: Built entirely with open-source technologies

🚀 Quick Start

Prerequisites

1. Node.js (v18 or higher)
2. Ollama - Install from ollama.ai
3. Whisper.cpp or OpenAI Whisper for speech recognition
4. Sox for audio recording (macOS: brew install sox)

Installation

1. Clone and navigate to the project:

cd ~/beast_mode

2. Install dependencies:

npm install

3. Install Ollama models:

ollama pull codellama
ollama pull llama2
ollama pull mistral

4. Install Whisper (choose one):

Option A: Whisper.cpp (Recommended)

git clone https://github.com/ggerganov/whisper.cpp
cd whisper.cpp
make
./models/download-ggml-model.sh base

Option B: OpenAI Whisper

pip install openai-whisper

5. Compile the extension:

npm run compile

Running Beast Mode

As VS Code Extension

1. Open the project in VS Code
2. Press F5 to launch the extension in a new VS Code window
3. In the new window, open a workspace/folder
4. Activate Beast Mode with Cmd+Shift+V (Mac) or Ctrl+Shift+V (Windows/Linux)

Commands

• Start Voice Input: Cmd+Shift+V or "Beast Mode: Start Voice Input"
• Select AI Model: "Beast Mode: Select AI Model"
• Execute Prompt: "Beast Mode: Execute Prompt"

🎯 Usage Examples

Voice Commands

Once voice input is active, say:

"Beast, create a new Python API that handles user authentication"

"Beast, refactor this function to use async/await"

"Beast, install express and create a REST API with CRUD operations"

"Beast, debug the error in app.js and fix it"

Text Commands

You can also use text-based prompts:

1. Open Command Palette (Cmd+Shift+P)
2. Select "Beast Mode: Execute Prompt"
3. Enter your command

⚙️ Configuration

Configure Beast Mode in VS Code settings:

{
  "beastMode.ollamaUrl": "http://localhost:11434",
  "beastMode.defaultModel": "codellama:latest",
  "beastMode.wakeWord": "beast",
  "beastMode.autoExecute": true,
  "beastMode.whisperModel": "base",
  "beastMode.voiceLanguage": "en"
}

Settings

• ollamaUrl: Ollama API endpoint (default: http://localhost:11434)
• defaultModel: Default Ollama model for coding tasks
• wakeWord: Wake word to activate voice commands (default: beast)
• autoExecute: Automatically execute commands without confirmation
• whisperModel: Whisper model size (tiny, base, small, medium, large)
• voiceLanguage: Language code for voice recognition (default: en)

🏗️ Architecture

beast_mode/
├── src/
│   ├── extension.ts              # Main extension entry point
│   ├── agent/
│   │   ├── aiAgentManager.ts     # AI agent orchestration
│   │   └── taskParser.ts         # Parse AI responses into tasks
│   ├── executor/
│   │   └── commandExecutor.ts    # Execute file/command operations
│   ├── ollama/
│   │   └── ollamaClient.ts       # Ollama API client
│   └── voice/
│       ├── voiceInputManager.ts  # Voice input coordination
│       └── whisperTranscriber.ts # Speech-to-text transcription
├── package.json
└── tsconfig.json

🔧 Development

Build

npm run compile

Watch Mode

npm run watch

Testing

npm test

🤝 Contributing

Contributions are welcome! This project is built on open-source principles.

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Submit a pull request

📋 Roadmap

• Enhanced wake word detection with custom training
• Multi-language support for voice commands
• Integration with additional AI providers
• Visual feedback for AI agent actions
• Conversation history and context management
• Plugin system for custom commands
• Web-based UI for configuration
• Docker containerization
• Cloud deployment options

🛠️ Troubleshooting

Ollama Connection Issues

Ensure Ollama is running:

ollama serve

Check available models:

ollama list

Voice Input Not Working

1. Verify Sox is installed: sox --version
2. Check microphone permissions in System Preferences
3. Test Whisper installation: whisper --help

Extension Not Loading

1. Check VS Code version compatibility (requires 1.85.0+)
2. Rebuild the extension: npm run compile
3. Check the Output panel for errors

📄 License

GNU Affero General Public License v3.0 (AGPL-3.0)

Beast Mode is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This means: if you run a modified version of Beast Mode over a network (e.g. as a hosted service), you must release the complete corresponding source code to users of that service. See the LICENSE file for full terms.

🙏 Acknowledgments

• Built on VS Code OSS
• Powered by Ollama
• Speech recognition via Whisper
• Inspired by Windsurf

📞 Support

For issues, questions, or suggestions:

• Open an issue on GitHub
• Check the troubleshooting section
• Review the configuration options

---

Beast Mode - Unleash the power of AI-driven development! 🦁