llama.vscode

Why we included this project

llama.vscode is a VS Code extension that runs LLM-assisted completion, chat, and agentic coding on local models through llama.cpp, so your source never leaves your machine. You accept inline suggestions with Tab, and the agent view can pull in MCP tools; smart context reuse keeps very large editing contexts responsive even on modest hardware, and the extension installs llama.cpp for you. It also works with external providers like OpenRouter, so you can switch between fully local and hosted models without changing interfaces. For teams that want modern assistant workflows but are uneasy about sending code to a cloud service, this is a practical middle ground.