A cybersecurity AI model that finds the way in before anyone else does. In two forms.
01Introduction
Kairo is a cybersecurity AI model from ALSoftworks. Kairo Blue defends; Kairo Red attacks systems you're authorised to test.
Kairo finds vulnerabilities, thinks like an attacker and helps you fix what it finds. It comes as two models: Kairo Blue for defence and Kairo Red for authorised offensive testing.
Why now. Attackers already use AI to find the weak spots. Kairo is built to find them first, and to show you exactly how to close them.
Early access. Kairo opens to security teams in waves. Every request is reviewed, and Kairo Red requests are checked by hand before anyone gets access.
02Two models
Same goal. Different permissions.
Both models look for the weak spots before someone else does. They differ in what they're allowed to do about them, and in who can use them.
- Built to
- Defend
- Main work
- Code and configuration review, threat hunting, alert triage
- Writes exploits
- No
- Access
- Security and engineering teams
- Delivers
- Findings with evidence and fixes
- Built to
- Attack, with permission
- Main work
- Penetration testing, attack-path chaining, proof of impact
- Writes exploits
- Only inside a verified scope
- Access
- Vetted security teams, reviewed by hand
- Delivers
- Attack paths with proof, steps and fixes
03How it works
How Kairo works
Kairo Blue defends
Kairo Blue reviews code and configuration, hunts for vulnerabilities, triages alerts and writes the fixes. It is built for defenders and doesn't produce working exploits.
Kairo Red attacks, on your terms
Kairo Red plans and runs penetration tests, chains small findings into real attack paths and proves the impact, only against systems you own or are authorised to test. Every Red request is reviewed by hand.
Fixes, not just flags
Every finding comes with the evidence, the impact and a concrete fix, so security teams spend their time closing holes instead of triaging noise.
04Safeguards
Built with limits.
An offensive security model is only as good as the limits around it. Kairo Red is built so it can't be pointed at systems nobody agreed to have tested.
- Scope first. Kairo Red only works against targets inside the scope a customer defines and verifies.
- People in the loop. Access to Kairo Red is limited to vetted security teams, and every request is reviewed by hand.
- A full record. Every action is logged, and a test can be paused or stopped at any time.
- Defence stays defence. Kairo Blue never writes working exploits or attack tooling.
05Availability
In early access.
Request Kairo Blue or Kairo Red on the Kairo site, and we'll email you when there's a place for your team. Benchmarks and research will appear in the Kairo lab as they're ready.