Open source · MIT

llama.cpp

The engine most local AI apps use to run open language models on ordinary computers.

130k stars on GitHub, Oct 2026

Why it's useful

Run open models like Qwen, Llama or Mistral on your own PC, including on AMD and Apple chips, with no API bill. Most local tools are built on it.

A computer with enough memory for the model you pick.

Image: ggml-org. llama.cpp belongs to ggml-org; we just think it's worth knowing about.

Models

by Ollama

The easy way to download and run open AI models on your own computer.

Video

by FlashML

A desktop app that runs the MiniMax H3 video model on an ordinary gaming PC.

Game Development

by ACE Studio and StepFun

Makes full songs or instrumentals from a text description, locally and quickly.