Private Copilot for Visual Studio Code

Private, Fast, Configurable AI Autocomplete

Private Copilot brings intelligent, low-latency AI code completion directly into your editor without sending your code to the cloud.

⚡ Key Features

• 🔒 100% Local-First Privacy — Source code stays on your machine (localOnly mode by default).
• 🚀 Sub-300ms Autocomplete — Instant keystroke debouncing, in-memory LRU request cache, and intelligent deduplication.
• 🧠 Multi-File Context Engine — Extracts imports, enclosing function/class scopes, and related workspace definitions with budget presets (fast, balanced, rich).
• 🔌 Universal Provider Support — Compatible with Ollama, LM Studio, vLLM, OpenAI, and custom OpenAI-compatible endpoints.
• 🎯 Fill-in-the-Middle (FIM) — Native support for Qwen 2.5 Coder, DeepSeek Coder, StarCoder2, and CodeLlama.
• 📊 Diagnostics Dashboard — Real-time Webview displaying latency percentiles (P50/P90/P95), acceptance rate, cache hit ratios, and language statistics.
• 🛠️ Status Bar Controls — Quick model switcher, latency indicator, and one-click connection tests.
• 🌐 15+ Programming Languages — TypeScript, JavaScript, Python, Go, Rust, Java, C, C++, C#, Ruby, PHP, Swift, Kotlin, and more.

🚀 Getting Started

1. Start your local model runtime (e.g. ollama run qwen2.5-coder:7b).
2. Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P) and run:

   Private Copilot: Setup Wizard
   

3. Follow the 5-step guided setup to connect to your endpoint and test connectivity.

⌨️ Common Commands

• Private Copilot: Setup Wizard — Guided configuration
• Private Copilot: Status Bar Menu — Interactive status menu
• Private Copilot: Toggle Enable/Disable — Turn autocomplete on or off
• Private Copilot: Quick Settings — Adjust settings without opening preferences
• Private Copilot: Show Diagnostics — Open real-time metrics dashboard
• Private Copilot: Clear Cache — Flush in-memory completion cache

📄 License

MIT