Free Tool Lets You Run AI Chatbots on Your Own Laptop

llama.cpp is a free piece of software that lets you run AI chatbots on your own device instead of renting them from a company over the internet. Think of it like the difference between streaming a movie and downloading it to your laptop — same movie, but you own the experience and nobody watches what you're doing. On September 23, the team behind it published a new version, labeled b11118, aimed mainly at making that experience faster and more reliable.

The headline benefit is privacy and money. When you use most AI chatbots, every question you type travels to a company's data center and gets stored somewhere. With llama.cpp, your words never leave your machine. There's also no subscription — you can chat as much as your computer can handle for free. The tradeoff is that you're limited by your own hardware, not a giant server farm.

This particular release is a technical one. Its main change involves improving how the software handles a specific kind of memory access for AI image and attention processing (a fix listed as 'hex-dma: introduce direct-mapped DMA cache'). The bigger story is the list of devices supported: Macs with Apple Silicon, Windows and Linux PCs, Android phones, and even phones with Snapdragon chips, where the AI runs on dedicated NPU hardware (a chip built specifically for AI tasks). That means the same free tool now runs almost anywhere you do.

The catch: this is a developer-focused build, not a polished consumer app. Installing it takes some comfort with downloading files and running commands, and it won't match the raw power of paid services like ChatGPT or Claude. Smaller AI models that fit on a laptop are less capable than the cloud giants. Still, for privacy-minded people or anyone tired of subscription fees, it's a genuinely useful option — and it keeps getting better every week.