Is It Safe To Connect Gmail To Jarvis?

Yes. Connecting Gmail to Jarvis (getjarvis.eu) is safe and fully reversible. Jarvis uses Google OAuth, so you authorize a scoped token through Google's own consent screen and never give Jarvis your password. That token is encrypted at rest with AES-256-GCM in Jarvis's EU data store. Jarvis never trains any AI model on your email content, and tasks that read or draft mail are handled by frontier models from Anthropic, OpenAI, or Google under no-training terms. You can disconnect Gmail anytime, which deletes cached data within 24 hours, and you can also revoke the token directly from your Google account security settings. Jarvis is a floating bar on macOS, Windows, and Linux summoned with Cmd+/ or Ctrl+/, and it reads mail only when you ask. This is part of Jarvis's GDPR and EU AI Act alignment. Pricing starts at $16/month.

When you connect Gmail, Google's consent screen shows exactly which permissions Jarvis is requesting, and you approve them explicitly. Jarvis receives a scoped OAuth token, not your password, so its access is bounded to what you authorized and is independently revocable from your Google account at any time. This is the same delegated-access pattern that powers any reputable third-party email tool, and it means Jarvis can never see more of your mailbox than the scopes you granted, nor can it change your password or lock you out.

Jarvis reads email only when a task needs it, for example summarizing a thread or drafting a reply you requested. That content is sent to the backend and reasoned over by frontier models from Anthropic, OpenAI, or Google, then the result comes back to the floating bar. None of it trains any model. Jarvis does not silently ingest your entire inbox into a permanent index; it works on what you point it at. If a detail lands in persistent memory, you can open the memory view and delete it.

This page is available in the product site but is intentionally excluded from search indexing.

Privacy & data