Top AI Product

llama.cpp shipped for three years with no website. You compiled it yourself or you didn’t use it. Now the ggml team put a product page in front of it — llama.app — and the whole thing collapses into curl -LsSf https://llama.app/install.sh | sh, or Brew, or Winget. HN gave it 221 points.

What it actually is

A local AI runtime, not a chat app. The installer drops a single llama binary that swallows the old zoo of tools (llama-server, llama-cli) behind one CLI. Hand-tuned kernels cover Apple Silicon, plain CPUs, RTX 3090 through 5090, H100, AMD Radeon, Intel Arc, Jetson. Qwen 3.6, Gemma 4, GPT-OSS run out of the box. The pitch: run frontier AI entirely on your machine. No API keys, no telemetry, no limits.

The part that matters for developers

llama-server exposes an OpenAI-compatible HTTP endpoint, so anything that speaks the OpenAI API — your app, your IDE plugin — points at localhost and works. And llama.app ships alongside Pi, a local coding agent that auto-discovers models you’ve already downloaded, no config file.

That last bit is the real move. Local Claude Code, free, MIT-licensed.

You Might Also Like

• Ollama mlx on Apple Silicon 1810 Tokens sec Prefill and the end of Llama cpp on mac
• Zcode Zhipu z ai glm 5 2 Coding Agent hit hns 1 Spot Chinas Open Answer to Claude Code
• Cloudrouter Gives Your ai Coding Agent its own Cloud Machine and Thats a big Deal
• Ggml Llama cpp Joins Hugging Face and Honestly it was Only a Matter of Time
• Claude Code Remote Control Just Turned my Phone Into a Coding Terminal and im Weirdly Into it