The problem
All teams lack access to agentic security products that run locally and improve without training on their data. By agentic, we mean software that can perform security work and help resolve what it finds instead of only creating another alert.
Today, teams are forced to choose. Periodic, human-led assessments leave gaps between reviews while software and attackers keep changing. Continuous AI products can move faster, but they often require teams to send source code and security findings to outside platforms, where third-party models can train on derivatives of that data.
Developers are left with more findings than time to fix them. CISOs piece together code scanning, offensive assessments, and vulnerability disclosure across multiple vendors, yet still cannot get continuous coverage while keeping sensitive data inside their environment. Teams need efficiency, security, and privacy together. Existing products make them choose.
Our mission
Vaktex exists to remove that tradeoff. We build agentic security products that run where the customer chooses and improve without training on customer data. They can operate on developer workstations, private servers, hardware appliances, and air-gapped networks.
A team should not have to send its code and findings to a third-party AI provider to get security software that stays current and works toward resolution. Our mission is to make continuous, private security available to every team.
What we build
Vaktex builds local agentic security products that continuously find, test, triage, and help resolve vulnerabilities without training on customer data.
- VScan runs local white-box assessments of source code and packages, then uses secure agents to help developers resolve vulnerabilities in their existing coding and code review workflows.
- Angrep continuously runs black-box and gray-box offensive assessments against public applications, networks, and sites.
- Meld manages public vulnerability disclosures, uses agents to triage reports, and adds human support where automation is not enough.
- VSecure provides secure local inference, manages model licensing and DRM. It runs on workstations, servers, appliances, and air-gapped networks.
Every service connects with VScan. Teams can keep code and security data inside their network, even in an air-gapped deployment, or optionally enable cloud metrics for a complete view of their security posture. Developers get help resolving findings, while CISOs get continuous oversight without handing sensitive data to third-party AI providers.
Why Vaktex is different
Most security products solve only one side of the problem: cloud automation that requires customer data, or local tools that do not keep improving. Vaktex is the first company to offer cyber models that are small enough to run locally, improve without training on customer data, and work toward resolution instead of stopping at detection.
Our models stay current through proprietary information streams, simulated vulnerable environments, and reinforcement learning. This means we do not benifit from training on your customer data. That is part of the problem we set out to solve: making agentic security continuous and private at the same time.
Where we started
Vaktex was founded in Texas by Eoin, Founder and CEO. The company grew from his years working across penetration testing, software development, machine learning, and product design. In 2026, that work became VScan and the first Vaktex models, Minisøk and Vakttårn. Today, Vaktex connects them with Angrep, Meld, and VSecure as one security platform.
We selected the name Vaktex as a blend of three concepts reflecting our company's mission and roots. “Vakt” is the Norsk term for a “security guard,” symbolizing both the detection and protection our services provide. “Vak” in Norsk means “awake” in English, describing vigilance against current threats through our ongoing model updates. Lastly, “Tex” is for Texas, where our product was created.