Sidekick AI - Your Private AI Coding Assistant

🚀 AI-powered coding, 100% offline. No cloud, no subscriptions, no data mining. Just pure, local AI assistance.

✨ Why Sidekick AI?

Unlike cloud-based AI assistants, Sidekick AI runs entirely on your machine. Your code never leaves your computer. Perfect for:

• 🔒 Enterprise environments with strict security policies
• 💼 Proprietary codebases that can't use cloud services
• 🌍 Offline development without internet dependency
• 🛡️ Privacy-conscious developers who value data ownership

🎯 Core Features

🤖 Smart Code Fixes

See an error? Click the lightbulb and let AI fix it. Sidekick analyzes your entire file context to provide intelligent fixes, not just simple corrections.

💬 Interactive AI Chat

Press Ctrl+Shift+A to open a GitHub Copilot-style chat panel. Ask questions, get explanations, or request code improvements.

🔍 Instant Code Explanations

Right-click selected code and choose "Explain Code" or press Ctrl+Shift+E to get a clear explanation. Perfect for understanding complex algorithms or unfamiliar codebases.

🔧 Intelligent Refactoring

Right-click selected code and choose "Refactor Code" to get AI-powered improvements. From performance optimizations to readability enhancements.

🧪 Automatic Test Generation

Generate comprehensive unit tests for your functions with a single click. Supports multiple testing frameworks.

⚡ Fast Inline Completions

Get intelligent code suggestions as you type. Trained on modern coding patterns, it understands your intent and provides relevant completions.

🚀 Quick Start (2 minutes)

Step 1: Install llama.cpp

Windows:

# Download from GitHub releases
https://github.com/ggerganov/llama.cpp/releases

Mac:

brew install llama.cpp

Linux:

git clone https://github.com/ggerganov/llama.cpp
cd llama.cpp && make

Step 2: Get a Model

Download our recommended model (1.5GB): Qwen 2.5 Coder - Optimized for code

Step 3: That's It!

Install Sidekick AI and it will:

• ✅ Auto-detect llama.cpp
• ✅ Guide you through model selection
• ✅ Start the AI server automatically
• ✅ Remember your settings

📊 Performance

• Response time: ~500ms for completions
• RAM usage: 2-4GB depending on model
• CPU only: No GPU required
• Token speed: 20-30 tokens/second on modern CPUs

🔒 Privacy Guarantee

{
  "telemetry": "NONE",
  "data_collection": "ZERO",
  "cloud_services": "DISABLED",
  "network_requests": "BLOCKED",
  "your_code": "STAYS_LOCAL"
}

🛠️ Configuration

Access settings with Ctrl+, and search "sidekick-ai":

{
  "sidekick-ai.modelPath": "path/to/your/model.gguf",
  "sidekick-ai.llamaPath": "auto-detected",
  "sidekick-ai.contextSize": 4096,
  "sidekick-ai.temperature": 0.7,
  "sidekick-ai.port": 8080
}

🐛 Troubleshooting

Github

https://github.com/Naveen-babu1/sidekick-ai

📈 Roadmap

• [ ] Streaming responses
• [ ] Multiple model support
• [ ] Custom prompt templates
• [ ] Project-wide refactoring
• [ ] Code documentation generation
• [ ] Git commit message generation

📜 License

Use it, modify it, share it freely!

Built with ❤️ for developers who value privacy and local-first AI

No account required. No API keys. No subscriptions. Just install and code.