Does Jarvis Keep My Documents Private When I Compare Them?

Yes. Jarvis (getjarvis.eu) is built so your documents stay private even when it reads, compares, or extracts from them. Jarvis is a screen-aware desktop AI assistant for macOS 12+, Windows 10+, and Linux that you trigger with Cmd+/ on macOS or Ctrl+/ on Windows. It encrypts every OAuth connector token, for Google Drive, Google Docs, Notion, and the rest, with AES-256-GCM, hosts all processing in the EU, and aligns with GDPR and the EU AI Act. Most importantly, it never trains on your prompts, screenshots, connector data, or memory, so a contract or report you analyzed today does not feed any model tomorrow. Its persistent memory is user-controlled: you can inspect, edit, or clear it. The underlying models, frontier models from Anthropic, OpenAI, and Google, run without your data being retained. Free to start, then $16/month.

The single biggest privacy concern with document AI is whether your files become training data. With Jarvis the answer is no: it never trains on your prompts, the screenshots it reads, your connector data, or its memory. When you compare two contracts or extract a table, that content is used only to produce your answer in the moment, not absorbed into any model. This is a hard commitment, not a setting you have to find and toggle, which means even routine document work carries no risk of your material resurfacing elsewhere.

Jarvis protects data in transit and at rest. OAuth tokens that grant access to Google Drive, Notion, Airtable, and your other apps are encrypted with AES-256-GCM, so even the keys to your accounts are not stored in the clear. All processing happens on GDPR-aligned infrastructure at rest, keeping your documents within European jurisdiction and aligned with GDPR and the EU AI Act. For European users and businesses handling regulated data, this EU-resident posture is often a requirement, not a nicety, and it applies to every document operation.

Docs & files