Ollama Lets You Run Local AI Models Inside ChatGPT Desktop

A new version of Ollama, a free tool that runs open-source AI models like Llama directly on your own computer, now lets you use those models inside the ChatGPT desktop app. Instead of being stuck paying for OpenAI's cloud models or juggling different apps, you can stay in ChatGPT while tapping into local models that work offline and keep your data private. Setup happens right from the Ollama app on a Mac, and it's designed to fit into your normal workflow.

So what does this actually mean for you? If you've been curious about AI models that don't send your conversations to a company's servers, this gives you a way to try them with a familiar interface. Local models can save you money on subscription fees, run even without internet, and give you more control over how your data is used. You can use them for drafting emails, brainstorming, or summarizing documents, all while keeping the polished ChatGPT experience. The update also claims better speed on Apple Silicon and fixes for handling images through compressed responses, so it's not just a new integration, but a smoother one.

The catch? This is still a pre-release version, which means it's not fully polished and might have bugs. It also requires the ChatGPT desktop app, the Ollama app, and currently works only on macOS. Running serious AI models locally demands a fairly powerful computer, so not every machine will handle it well. But even with those limits, this is a step toward a future where you can mix and match the best AI tools, taking cloud models when convenient and private local models when it matters.