Privacy should be architecture, not fine print.
The privacy law.
Seven durable rules govern what the organizational brain remembers, who can see it, and where the boundary sits.
Information you choose to share.
When you request beta access, we receive the email address, use case, and setup details you enter. We use that information to evaluate the request, place you in an appropriate cohort, and contact you about access. We do not sell it.
The large sentence is the promise. The text beneath it explains the operating rule that makes the promise concrete.
- R1
What you ask is yours. What you tell the brain belongs to the organization.
Reads are private; writes are visible within their scope. Queries and block reads are member-private by default. Writes, statements, and attestations are accountable organizational acts.
- R2
Patterns surface only when the crowd is large enough to protect the individual.
Aggregates are k-anonymized at k ≥ 5. Topic-level signals appear only above that threshold, preventing small teams from being de-anonymized by arithmetic.
- R3
The organization sees what is missing—not who went looking.
A knowledge-gap signal contains a topic and a count, never the identity of the person who queried it.
- R4
Curiosity about restricted knowledge stays protected.
Access requests are visible only to the permission holder who can grant them, not to feeds or broad organizational views.
- R5
Agents work in the open; humans think in private.
Agent cognitive events are organizational work product by default. An agent operating inside a person's private scope inherits that privacy boundary.
- R6
If an organization audits, everyone knows.
Audit mode is off by default and is never silent. When enabled, every member is shown that the organization audits query logs.
- R7
Your identity fades from query logs after 30 days.
Raw identified query logs roll into anonymous aggregates after 30 days. Outcome events keep their authors because attestations and recorded work are accountable acts.
Ask what we hold. Correct it. Delete it.
For a privacy request or question, contact virajsharma@kaeva.app.