NoaxAI 🌠 - Your Local AI Coding Assistant

Your AI coding companion that runs completely offline!

Run Hugging Face models locally in VS Code! Code with AI assistance completely offline using runtime.

NoaxAI brings the power of large language models directly into your editor - no cloud services required. Download models once, then work completely offline with AI-powered coding assistance.

Why NoaxAI? 🤔

• 🔒 Privacy First: All AI operations run locally on your machine
• 💨 Fast & Efficient: Uses optimized models for quick responses
• 🌐 Offline Ready: No internet needed after initial model download
• 🎯 Code Focused: Built specifically for programming workflows
• 💰 Free Forever: No subscriptions or API costs

Features

🤖 Enhanced Model Management

• Explore and manage AI models through the "Explore Models" panel
• Search models by name, size, or description
• Models are sorted by size for easy selection
• Visual indicators for installed and selected models
• Real-time download progress with progress bar
• Models range from 4K to 3B parameters
• Switch between different models instantly

💬 Advanced AI Chat

• Interactive chat interface with AI assistant
• Multi-file context support - attach multiple files to your questions
• Stream responses in real-time
• Export generated code directly to files
• Quick action suggestions and code examples
• Organized welcome screen with common tasks
• Stop generation at any time
• Visual file attachment system
• Rich markdown and code block support

🛠️ Code Operations

• Refactor Code: Improve code quality and efficiency
• Generate Unit Tests: Automatically create tests for your code
• Fix Bugs: Get suggestions for bug fixes
• Document Code: Generate comprehensive documentation
• Custom Operations: Use inline editing for custom code transformations
• Right-Click Menu: Quick access to NoaxAI features
• Context-Aware: Language-specific suggestions

📝 Documentation Generation

• Generate comprehensive documentation for entire folders
• Create user guides and technical documentation
• Generate ASCII diagrams for code structure
• Export to markdown format
• Support for mkdocs configuration
• Maintains folder structure in output
• Customizable site name and description

🖥️ REPL Integration

• AI-powered command-line helper
• Get suggestions for terminal commands
• Execute suggested commands directly
• Interactive command selection
• Real-time command output display
• Multiple command suggestions

Installation

1. Install NoaxAI extension from VS Code marketplace
2. Open VS Code
3. Open the "Explore Models" panel in the sidebar
4. Browse or search for a model that fits your needs
5. Click on a model to install it locally
6. Wait for the model to download (progress bar will show status)
7. The model will be automatically selected after installation

Usage

Model Selection & Search

1. Open the "Explore Models" panel
2. Use the search icon to find models by name, size, or description
3. Click on the cloud icon to install a model
4. Installed models show a checkmark
5. Selected model shows "Selected ✓" indicator
6. Models are sorted by size for easy selection

Enhanced Chat Interface

1. Click the chat icon in the status bar or use command palette
2. Use 📎 to attach relevant workspace files
3. Toggle "Export" to save generated code to files
4. Real-time streaming responses
5. Stop generation at any time
6. Organized suggestions for common tasks
7. Visual file context display

Code Operations

Select code and right-click to access:

• NoaxAI: Refactor Code
• NoaxAI: Generate Unit Test
• NoaxAI: Fix Bug
• NoaxAI: Edit Inline

Documentation Generation

1. Right-click on a folder in the explorer
2. Select NoaxAI: Document This Folder
3. Enter site name and description
4. Documentation will be generated in markdown format

REPL Helper

1. Open command palette
2. Type NoaxAI: REPL
3. Ask for command suggestions
4. Select and execute suggested commands

Keyboard Shortcuts

• Ctrl+N: Apply last AI suggestion when text is selected

Requirements

• VS Code ^1.94.0
• Internet connection for initial model download
• Sufficient disk space for local models

Extension Settings

This extension contributes the following settings:

• NoaxAI.cacheDirPath: Path to store downloaded models
• NoaxAI.selectedModel: Currently selected model for operations

Known Issues

• Initial model download may take time depending on your internet connection
• Model switching may have a brief delay

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

MIT

Credits

• Models provided by Hugging Face
• Runtime for model execution
• Various open-source contributors

Performance 📊

Model	Size	Speed	Memory Usage
SmolLM2-135M	135MB	Fast	~300MB RAM
Llama-3.2-1B	1.1GB	Standard	~2GB RAM

🌟 Request New Models

Need a specific model? We're constantly expanding our model support!

How to Request:

1. Check if the model is compatible
2. Create a Model Request on GitHub
3. Provide the following information:
   • Model name and Hugging Face link
   • Use case/scenario
   • Expected benefits

Current model priorities:

• 🚀 Small & fast models (< 500MB)
• 💻 Code-specialized models
• 🔧 Task-specific models

Request a Model Now →

Support & Community 👥

• Star the repo to show support! ⭐