Free AI Tool llama.cpp Just Got a Fix for Your Browser

llama.cpp is a free, open-source program that lets you run AI chatbots on your own computer — no monthly fee, no internet connection, no data leaving your machine. It's wildly popular: 129,000 people have "starred" it on GitHub (the site's version of a like), and 23,400 have copied it to build their own projects. Think of it as the engine under the hood of many private AI apps.

This week's update, labeled b11037, is a small maintenance release. The fix targets the browser version of the tool, which uses WebGPU (a way for web pages to borrow your graphics card for heavy math). Specifically, it corrects a bug in how the software "gets rows" — pulling out chunks of text data while the AI writes an answer. Left unfixed, this could produce garbled output or crash the page.

The release also ships ready-made versions for almost every computer: Mac, Windows, Linux, Android, iOS, plus special builds for Nvidia, AMD, and Intel graphics chips. That matters, because installing AI software is usually the hardest part. Here you download a file and run it — the difference between assembling furniture and having it delivered to your door.

The catch: this is a bug fix, not a new ability. It won't make your AI noticeably smarter or faster. And running a good AI model locally still demands a reasonably modern machine with enough memory; a five-year-old budget laptop will struggle. The browser version stays the least capable option. For the full experience, the installed app is still the way to go.