The privacy architecture behind Arcana

Search is personal data

What you search for is a precise map of what you are thinking about, worrying about, researching, and deciding. Search history is among the most sensitive data a person generates. Most search engines monetize it. Arcana does not.

This is not a marketing claim. It is a structural choice that shapes how we built the product from the ground up.

What we do not collect

Arcana does not log your queries for advertising purposes. We do not build user profiles. We do not sell data to third parties. We do not use your searches to train our models. The query you submit to Arcana exists to generate your answer. Once that answer is delivered, your query is not retained for any purpose beyond the session.

Session context vs. persistent tracking

Arcana maintains context within a session so that follow-up questions make sense. This is functionally necessary: without short-term context, every message would require you to re-explain your situation. But session context is distinct from persistent tracking. When your session ends, that context does not feed into a long-term profile.

For users who enable search history, that history is stored under their account and is accessible only to them. It can be paused or deleted at any time.

Enterprise controls

Enterprise customers have additional controls: configurable data retention policies, audit logs, and the ability to restrict history features entirely for their organization. For teams handling sensitive research, these controls provide the compliance layer that individual privacy settings alone cannot.

The privacy default

We made privacy the default rather than the opt-in because most users do not change default settings. A privacy-respecting product that requires users to find and activate privacy settings is not actually privacy-respecting. The protection should be there from the first search.