Nexlyt AI Code Agent

🤖 AI-powered coding assistant for VS Code supporting local Ollama and OpenAI.

✨ Features

🗣️ Chat Interface

• Interactive AI chat panel with modern UI
• Streaming responses with typing effect
• Persistent chat history across sessions
• Quick action buttons for common tasks
• Export chat history to markdown

🚀 AI Actions

• Explain Code - Get detailed explanations of selected code
• Refactor Code - Improve code structure and readability
• Optimize Code - Enhance performance and efficiency
• Generate Tests - Create comprehensive unit tests
• Generate Documentation - Add JSDoc/docstring comments
• Find Bugs - Detect potential issues and security problems

⚙️ Dual AI Support

• Ollama (Local) - Run AI models locally for privacy
  • Auto-detects available models
  • Model auto-pull if missing
  • Supports mistral, llama3, codellama, and more
• OpenAI (Cloud) - Use GPT models via API
  • Secure API key storage
  • GPT-3.5-turbo and GPT-4 support

⌨️ Quick Access

• Keyboard shortcuts for common actions:
  • Ctrl+Alt+C - Open AI Chat
  • Ctrl+Alt+E - Explain Selection
  • Ctrl+Alt+R - Refactor Selection
  • Ctrl+Alt+T - Generate Tests
• Context menu integration - Right-click selected code
• Command palette - All commands available via Ctrl+Shift+P

🔧 Advanced Configuration

• Custom Ollama URL - Connect to remote Ollama instances
• Temperature control - Adjust AI creativity (0.0-2.0)
• Max tokens - Control response length (512-32768)
• Context lines - Set surrounding code context (10-200)
• Auto-insert - Automatically apply AI suggestions
• Streaming mode - Enable/disable typing effect

🚀 Quick Start

1. Install Extension

• Download and install the .vsix file, or
• Install from VS Code marketplace (coming soon)

2. Setup AI Provider

Option A: Ollama (Local, Recommended)

1. Install Ollama: Download from ollama.com
2. Pull a model: Run ollama pull mistral in terminal
3. Verify: Extension will auto-detect running Ollama

Option B: OpenAI (Cloud)

1. Get API key: Visit platform.openai.com
2. Set key: Run command Nexlyt: Set OpenAI API Key
3. Switch provider: Set nexlyt.apiProvider to "openai"

3. Start Coding with AI

1. Open chat: Press Ctrl+Alt+C or run Nexlyt: Open AI Chat
2. Select code and right-click for AI actions
3. Use slash commands: /explain, /refactor, /optimize, /test

📖 Usage Examples

Chat Commands

/explain - Explain the selected code
/refactor - Refactor selected code for better structure  
/optimize - Optimize code for better performance
/test - Generate unit tests for selected code
/document - Generate documentation comments
/bugs - Find potential bugs and issues

Context Menu Actions

1. Select code in editor
2. Right-click → Choose AI action
3. View results in chat panel
4. Click "Insert" to apply suggestions

Keyboard Shortcuts

• Select code and press shortcuts for instant AI analysis
• Results appear in chat panel with insert button

⚙️ Configuration

Open VS Code Settings (Ctrl+,) and search for "Nexlyt":

Basic Settings

• nexlyt.apiProvider - Choose "ollama" or "openai"
• nexlyt.ollamaModel - Set Ollama model (default: "mistral")
• nexlyt.ollamaUrl - Ollama server URL (default: "http://localhost:11434")

Advanced Settings

• nexlyt.maxTokens - Maximum response length (default: 4096)
• nexlyt.temperature - Response creativity 0.0-2.0 (default: 0.7)
• nexlyt.contextLines - Code context lines (default: 50)
• nexlyt.autoInsert - Auto-apply suggestions (default: false)
• nexlyt.streamingMode - Enable typing effect (default: true)
• nexlyt.saveHistory - Persist chat history (default: true)

🛠️ Development

Build from Source

git clone <repository>
cd nexlyt-ai-code-agent
npm install
npm run compile

Debug Extension

1. Open in VS Code
2. Press F5 to launch Extension Development Host
3. Test in new window

Package Extension

npm install -g vsce
vsce package

🔒 Privacy & Security

Ollama (Local)

• Fully private - All processing on your machine
• No data sent to external servers
• Offline capable - Works without internet

OpenAI (Cloud)

• API key encrypted in VS Code secret storage
• Code sent to OpenAI for processing
• Standard OpenAI privacy policy applies

🤝 Contributing

1. Fork repository
2. Create feature branch
3. Make changes and test
4. Submit pull request

📄 License

MIT License - see LICENSE file for details

🆘 Support

• Issues: Report bugs on GitHub
• Discussions: Feature requests and questions
• Wiki: Detailed documentation and guides

Made with ❤️ by Nexlyt Team