An AI defense systemin two parts.

Tacit, a streaming model, reads security events as they happen. AI agents investigate what it flags and propose a response, which runs only when your rules allow it or a person approves.

How it works

The two parts

Tacit

A streaming model, in development.

Tacit reads security events (processes, credentials, network and resource activity) as they happen. It keeps context from one event to the next, so it can notice a pattern that no single event shows.

CoordinatorInvestigationThreat HunterDefenderForensicsRecoveryVerificationone incident

The agents

A team of AI agents, in research.

Specialist agents work on one incident together. They gather evidence, test explanations, plan a response and check that it worked. This layer is still a research project and is not part of Tacit yet.

Why we built it in two parts