LLM models

Models power chat and question-answering in OrbySystems. You can change which one you use at any time in Settings → Models.

The OrbySystems models

OrbySystems Small, OrbySystems Medium and OrbySystems Large are the models OrbySystems is built and tested for. They run entirely on your own computer, so nothing you ask, and none of your data, leaves it. Hardware requirements says which one suits your computer.

Other models

You can choose a different model instead: a larger open-source model that runs on your computer, a model you run yourself with software such as LM Studio, or a cloud model from an AI provider. OrbySystems isn't tested with these, so its answers can be less reliable.

Cloud models

A cloud model runs on someone else's servers: the state-of-the-art models from the leading AI providers. Using one means sending it what you'd otherwise keep on your computer: your questions, and the project data it reads to answer them.

That provider's terms of service govern what they say they'll do with what you send, but terms can change, and by the time they do, your data is already with them. It's also subject to the same risks as any data held by a third party: a leak or breach, whether from an outside attacker or someone inside the company. Weigh that before choosing a cloud model. The AI Model card in the sidebar always shows where the model you're using runs; see Models and your data.

Choosing and testing a model

See Models for how to choose one, and Model Test for checking that a model works well before relying on it.