Draft on sensitive data with the network off
Prove that nothing leaves your machine — work fully offline
Jan's whole point is privacy: because the model runs locally, you can use it on unpublished results, draft manuscripts, or other sensitive material that you would never paste into a cloud chatbot. The most convincing way to see this is to turn your network off entirely and watch Jan keep working. This is the tool for the work you cannot send to the cloud — it runs on a plane, in a no-internet lab, and never transmits your data.
- 1 Confirm a local model is loaded from lesson 00. Start a new chat and send a quick test prompt to be sure it responds.
- 2 Turn your network off — switch off Wi-Fi or unplug the network cable. Jan does not need a connection to run a local model.
- 3 Paste a piece of your real, sensitive work — for example a draft methods section or a summary of unpublished results — and ask:
Tighten this paragraph and flag anything that is unclear or unsupported. - 4 Read the reply. It arrives even with the network off, because the model is running on your own machine.
- 5 Reflect on what just happened: the unpublished text never travelled anywhere. With the network off there was no path for it to leave.
Jan produced a useful reply about your sensitive text while your machine had no network connection at all.
Build an offline drafting workflow for one real task
Turn the offline capability into a habit you can rely on for a recurring private task in your research.
Pick one genuinely private task you do regularly (drafting notes on unpublished results, summarising a confidential dataset's findings, rewriting a sensitive section) and run it end to end with the network off.
- 1 Choose a real, private task — not a toy example.
- 2 Turn the network off and complete the task in Jan from start to finish.
- 3 Note where the local model was good enough and where you would still want a stronger model.
- 4 Write one sentence describing when you will reach for offline Jan instead of a cloud tool.
A short note: the private task you ran offline and your one-sentence rule for when to use local Jan.