The Free AI App That Lives on Your Laptop Just Got Faster

Most AI chatbots you use live on somebody else's computers. You type, your words travel to a data center, and an answer comes back. llama.cpp does the opposite: it squeezes a full AI model onto your own laptop, phone, or desktop so the whole conversation happens on your hardware. Nothing you type leaves your machine, and nobody charges you a monthly fee. It's the project that quietly made "running AI at home" possible for ordinary people, and it has been downloaded and starred by hundreds of thousands of users.

The September 12 update, labeled b10931, adds a cache to the app's interface. Think of it like a bookmark that saves your place. Instead of redoing the same work every time you open the app or switch chats, it remembers what it already figured out. The practical result is less waiting and less strain on your computer. The same release also ships ready-made versions for Mac, Windows, Linux, Android, and iOS, including builds tuned for different graphics chips — so there's likely a download that fits your machine.

Why should you care if you're not a programmer? Because this is the free alternative to paying $20 a month for AI. If you have a reasonably modern laptop, you can chat with an AI that never sends your notes, contracts, medical questions, or diary entries to a corporation. That privacy angle matters more every year as people paste sensitive material into chatbots without thinking.

The catch is real, though. llama.cpp is a volunteer project, not a slick consumer app. Setup can involve copying commands into a terminal window, and the quality of the AI you get depends on your computer's memory and chip. A cache makes it faster; it doesn't make it easier. If you're curious and patient, it's a genuinely free path to private AI.