Questions people ask

Is it really free? For individuals, yes. For companies it is free under US$5 million of annual revenue, and a commercial license is needed above that (see LICENSING.md). There are no per-question fees either way; running it costs the electricity your computer uses.

Is it as smart as the big online assistants? No, and it doesn't pretend to be. The small models are good for drafting, summarising, explaining and brainstorming. The larger "standard" model is noticeably better. Its main advantage is privacy and zero running cost.

Does it need the internet? Not once it is installed and has a model.

Is my text sent anywhere? No. See PRIVACY.md.

Which model should I pick? Let OwnLLM choose (ownllm start picks the best one that fits your memory), or run ownllm models.

It's slow. Try a smaller model (ownllm start tiny), close other apps, and make sure the computer isn't low on free memory. ownllm doctor will point out problems.

Can I use it from other apps? Yes. After ownllm start, the API at http://127.0.0.1:8765/v1 speaks the common chat format; use the key from ownllm key.

Can other computers on my network use it? Not by design. If you need that, put a properly secured gateway in front of it that you control.