Run the assistant on your own computer
Run the assistant or a summary on a model that lives on your own computer. Use this when the content is sensitive, when you are working offline, or simply when you prefer your own model. Nothing is sent to your organization’s servers or to any external provider.
Before you start
Section titled “Before you start”- The model runs on your computer, so your computer does the work. A newer machine with more memory answers faster.
- A local model must be installed on your computer (the Ollama app, with at least one model downloaded). This is a one-time, per-computer setup — ask your IT team if it is not installed yet.
- Open the Assistant (bottom-right), or the summary card at the top of a topic.
- Set AI to Local (Ollama).
- Choose a model from the list. The list shows the models installed on your computer.
- Ask your question, or select the summary action.
Result
Section titled “Result”The answer is produced by the model on your computer. The prompt never leaves your machine — not even to your organization’s servers — and no other person can reach your model.
To go back to the shared model, set AI back to Server. Your choice is remembered for next time, per person.