Free App That Runs ChatGPT-Style AI on Your Own Laptop Gets a Tune-Up

Meet llama.cpp — a free, open-source program that lets you run AI chatbots directly on your own computer instead of paying a company to run them for you in a data centre. It's wildly popular with tinkerers: the project has been starred (bookmarked) nearly 128,000 times on GitHub, the website where code lives. Think of it as the difference between growing herbs on your windowsill and buying them at the shop every week.

The update released on September 16 is version b10993, and honestly, it's a housekeeping release. The team adjusted how the program uses your computer's processor (the main brain chip), and set up automated testing so future versions break less often. Nothing here changes how the app feels to use. But there's a genuinely useful part: the project published pre-built download packages for a long list of devices — Windows and Mac laptops, Android phones, Linux machines, and versions tuned for graphics cards from Nvidia, AMD and Intel.

That matters because 'pre-built' means you don't have to be a programmer. Normally, free software like this arrives as raw ingredients you must assemble yourself — a process that can take an afternoon and a lot of patience. Ready-made packages mean you download a file, double-click, and you're running an AI assistant.

Why would you bother when ChatGPT is free? Privacy and money. When you use a cloud chatbot, your questions travel to a company's servers. When you run one locally, nothing leaves your laptop — useful for work documents, medical questions, or anything you'd rather not hand over. You also stop worrying about monthly fees or usage limits. The honest catch: local AI is smaller and less clever than the big cloud models, and it needs a reasonably modern machine. This update doesn't fix that — it just makes the free, private option a little easier to reach.