Free Tool That Runs AI on Your Own PC Gets Windows Fix

A free piece of software called llama.cpp released a small update on September 28. If you've never heard of it, here's the short version: it's one of the most popular tools in the world for running AI chatbots on your own computer rather than sending your questions to a company's servers. It has been starred, or bookmarked, by roughly 130,000 people on GitHub — the site where programmers share code — which makes it one of the most-watched free AI projects around.

So what actually changed? Almost nothing, on purpose. The update fixes a bug where Windows computers were saving downloaded AI models in the wrong place. Models are the large files that give an AI its smarts, and services like Hugging Face are where you get them. When the storage path was wrong, Windows users could end up confused about where their files went, or downloading the same large file twice and wasting disk space and time. One line of code, essentially, but it removes a small daily annoyance.

The bigger story is the pattern. While the headlines focus on paid chatbots and billion-dollar data centers, a quieter movement is making capable AI free and private. Running an AI on your own machine means your questions never leave your laptop — no subscription, no monthly bill, no company reading your chats. That is genuinely appealing for anyone handling sensitive work, medical notes, or legal documents.

The catch: this is a maintenance release, not a magic upgrade, so it won't make your computer faster or smarter overnight. And running AI locally still demands a reasonably modern machine, a few gigabytes of free storage, and enough patience to follow setup instructions. This update just makes one step of that process less frustrating for Windows owners. If you already use llama.cpp, grab it. If you don't, it's a reminder that free, private AI is closer than you think.