Free Tool Runs AI Chatbots on Your Laptop, No Subscription Needed

A popular free software project called llama.cpp released a new version this week, labelled b11237. In plain terms, llama.cpp lets you run AI chatbots — the same kind of technology behind ChatGPT — directly on your own computer or phone, instead of paying a company to use theirs. This particular update fixes a small technical issue in how the software talks to certain Intel chips, so the AI runs more smoothly on more machines. It is a plumbing repair, not a headline feature, but it keeps the whole system working across a huge range of devices.

Why should you care? Normally, using a capable AI assistant means a monthly subscription and sending your private conversations, documents and questions to a company's servers. llama.cpp flips that. The AI model sits on your device, works without an internet connection, costs nothing, and nothing you type leaves your laptop. It supports chips from Apple, Nvidia, AMD, Qualcomm and Intel, and runs on macOS, Windows, Linux, Android and iOS.

The numbers tell the story of how big this has become. The project has roughly 130,000 stars and 23,900 forks on GitHub — shorthand for a massive community of volunteers improving it. This one release includes builds for Macs, iPhones, Windows PCs, Android phones, and even Chinese server chips. That breadth means AI is quietly becoming a commodity you can run anywhere, rather than a service you rent from a handful of giants.

The honest catch: this is a pre-release aimed at developers, not a polished app. The change itself is narrow and most people will never notice it. If you want the benefits without the technical work, you would use a friendlier app built on top of this engine, such as LM Studio or Ollama. And the biggest AI models still need a reasonably powerful machine — an old, cheap laptop can only run the smaller ones.