Official home for llama.cpp

llama.app is the official home for llama.cpp, a tool for running frontier AI models entirely on your local machine with no API keys, telemetry, or data leaving your device. It supports any hardware from laptops to clusters with hand-tuned kernels for GPUs and CPUs.

The site highlights integration with a local coding agent via the Pi tool and pi-llama plugin, and showcases supported models including Qwen 3.6, Gemma 4, GPT-OSS, and Gemma 3.

Questions this post answers

How do I set up llama.cpp with a local coding agent using Pi?

Run llama serve to start a local model server, then install the pi-llama plugin with pi install git:github.com/huggingface/pi-llama, and finally run pi. Pi automatically discovers the local model with no config or API keys required. All files stay on your machine and requests never leave it. Developers building local AI coding workflows track llama.cpp updates and integrations on daily.dev.