Skip to main content
Platform
Platform Overview How It Works
Solutions
AppSec Teams Platform Engineering Pricing Blog
Sign In Request Early Access
About Cybret AI

Built by practitioners who got tired of triage theater

We built Cybret AI because we watched skilled security engineers burn hours triaging findings that could never be exploited. We wanted to change that ratio.

Our mission

Security teams are buried under volume. Scanners surface thousands of findings a week. Most of them point to code that production workloads never actually execute. The real risk stays hidden inside that noise. Cybret AI maps your application's actual execution paths and tells you which vulnerabilities lie on routes an attacker can traverse from an entry point. Not which packages have a CVE. Not which patterns match a rule. Which specific code paths in your specific deployment are genuinely exposed.

Team

The people building it

A focused team with backgrounds in application security research, distributed systems, and developer tooling.

Adrian De Gendt

Adrian De Gendt

Co-founder & CEO

AppSec architect with experience designing triage workflows for production microservice environments. Focused on the gap between scanner volume and actionable signal.

Mikael Sundqvist

Mikael Sundqvist

Co-founder & CTO

Graph-traversal engineer who has spent years building static analysis and security infrastructure pipelines. Designed the call-graph engine at the core of Cybret AI.

Selin Kaya

Selin Kaya

Head of Security Research

Vulnerability researcher with a background in exploit path analysis and threat modeling. Brings an attacker-perspective lens to how Cybret AI classifies and surfaces exposure.

Origin

Started from frustration

In 2023, Adrian was running security triage sessions where the same pattern kept appearing: the majority of flagged findings lived in code paths that production traffic never reached. Developer time was being consumed patching things that could not be exploited, while the findings that actually mattered stayed buried in the same queue.

Mikael had been building call-graph tooling for performance profiling and recognized that the same graph infrastructure could answer the reachability question for security. They joined forces, brought Selin in to apply a vulnerability research perspective to the classification model, and started building toward the early-access program currently underway.

Join the early-access group

We are working closely with a small set of security and platform teams. If you want to shape the roadmap, apply now.