Copilot for llama-server LLMs

Local AI-powered coding assistance in VS Code via llama-server. Chat, inline completions, and cursor-rules integration — all running on your own hardware.

Note: This project has no affiliation with llama.cpp or its maintainers.

Packages

This is a monorepo containing two VS Code extensions and a shared library:

Package	Description
packages/workspace	Copilot for llama-server LLMs — chat provider, endpoint configuration, inline completions, cursor rules. Published as the primary Marketplace extension.
packages/ui	Llama Copilot — Server Manager — downloads, runs, and manages a local llama-server instance. Installed automatically as part of the workspace extension pack.
packages/shared	Shared types, constants, and utilities used by both extensions.

Build

npm install
npm run compile

Test

npm test

Package

Build .vsix files for both extensions:

npm run package

This produces packages/workspace/*.vsix and packages/ui/*.vsix.

Links

• llama.cpp GitHub
• llama-server Documentation