Ollama's New Update Brings Local AI to ChatGPT Desktop
Run private, free AI models right inside your ChatGPT app.
If you've ever used ChatGPT, you know how convenient it is. But you might also worry about privacy, or want to try AI models that aren't owned by a big company. That's where Ollama comes in. Think of it as a friendly app that lets you download and run hundreds of free, open-source AI models directly on your own computer — no cloud, no data leaving your machine.
Now, with version 0.34.0, Ollama models can work directly inside ChatGPT Desktop. Right now it's only on MacOS, but the setup reportedly happens right from the Ollama app. For everyday users, this means you can keep using the ChatGPT interface you already know — the chat window, the buttons, the look and feel — but behind the scenes, the actual AI answering you is one you control. That's a big deal for privacy, for offline use, and for anyone who wants more options beyond the big corporate AI services.
The update also brings other improvements: faster structured outputs on Apple Silicon (so your Mac runs certain AI tasks more smoothly), better support for OpenAI-compatible tools, and a fix that makes images work correctly in compressed responses.
One honest catch: this is a pre-release version, meaning it's still getting tested. So if you're not comfortable with occasional bugs, you might want to wait for the stable release. But if you're curious about running open models with a familiar interface, this is a great step forward.
- You can now use free, open-source AI models inside the ChatGPT Desktop app on Mac.
- Your chats run locally on your computer, which means more privacy and the ability to work offline.
- This is a pre-release version, so expect some bugs before the official rollout.
Why It Matters
This gives everyone a rare chance to use the ChatGPT interface with private, local AI — not just big tech's cloud.