ggml-org/llama.vscode

VS Code extension for LLM-assisted code/text completion

What it solves

It provides a way to integrate local Large Language Models (LLMs) directly into VS Code for real-time text completion, AI chat, and agentic coding, reducing reliance on cloud-based AI services.

How it works

The extension connects to a llama.cpp server running on the user's machine. It uses Fill-In-the-Middle (FIM) compatible models to provide auto-suggestions as the user types. It also includes a "Llama Agent" that can interact with the codebase using internal tools, custom JavaScript tools, and Model Context Protocol (MCP) servers installed in VS Code.

Who it’s for

Developers who want local AI-assisted coding tools that work offline, maintain privacy, and can run on consumer-grade hardware.

Highlights

  • Local FIM Completions: Real-time auto-suggestions with customizable acceptance keys (Tab, Shift+Tab).
  • Llama Agent: An agentic coding assistant with access to internal tools and MCP server tools.
  • Smart Context Management: Supports large contexts on low-end hardware through smart context reuse.
  • Model Management: Ability to search, download, and organize models from Hugging Face into "environments" (groups of models).
  • External Integration: Includes a Telegram bot for remote access to agents and deep-link support to open VS Code prompts.

Related

  • Project
    ggml-org/llama.vim

    A Vim/Neovim plugin that provides local LLM-assisted text completion and instruction-based editing via a llama.cpp server.

  • Project
    PawanOsman/OpenCursor

    An open-source AI coding agent for VS Code that supports local-first execution via llama.cpp and Ollama, featuring semantic codebase search and integrated developer tools.

  • Project
    s-kostyaev/ellama

    An LLM assistant for Emacs that provides chat, agentic coding workflows, and deep context integration across buffers and files.

  • Project
    yaroslavyaroslav/OpenAI-sublime-text

    An AI assistant plugin for Sublime Text that integrates multiple LLM providers, including local models via Ollama and llama.cpp, to provide chat and inline code assistance.

  • Dispatch
    llama.cpp and llama.app: Local LLM Inference and Ecosystem

    llama.cpp and its official home llama.app provide high-performance, hardware-agnostic local LLM inference, now featuring a simplified 'llama serve' command and integration with local coding agents like Pi.