Our mission
Software that defends itself.
A Swiss applied AI cybersecurity research lab, building autonomous systems that find and fix unknown vulnerabilities in the software companies run and depend on.
As more software gets written by AI, quarterly pentests become less useful by the day. If your security testing happens because compliance needs a report or legal wants a checkbox, it’s not going to end well.
Scanners run constantly, but create so much noise that teams learn to ignore them. Bug bounties incentivize hunters to report vulnerabilities to the good guys first, but only find what someone happens to look for. And a new wave of “always-on AI pentesting” companies is turning pentests into SaaS: another dashboard, another queue of findings to triage, another security tool your team has to learn and operate.
But running the old model more does not fix the model itself.
AI is removing the human bottleneck from vulnerability research. Finding new attack paths is no longer limited by how many researchers you can hire or how many hours they can spend looking. Software can now be attacked at machine speed, and vulnerabilities can be discovered, weaponized, and exploited while defenders are still waiting for the next scan, pentest, or patch cycle.
Cybersecurity cannot keep operating as a sequence of checks.
At 0sec, we’re building for what comes next: autonomous security systems that continuously attack, and fix your software before someone else does. A security team that operates itself, and gets out of your way as much as possible.
Core values.
We prove it before we report it
A blind agent re-exploits every finding without knowing how the first one worked. What it can’t reproduce never reaches you.
We work in public where we can
Open-source findings go upstream under coordinated disclosure, in public archives, under our name. Client work stays private.
We believe in open-source
Published under MIT OR Apache-2.0. Read it, run it on your own code, check that it does what we say.