Security researchers building for AI
RAXE was founded on two beliefs: AI threats are real and accelerating, and the security industry's instinct to inspect everything in the cloud is fundamentally wrong for AI, where prompts contain the most sensitive data in the system.
Where we came from
We started where the threats are—inside the application, on the developer's machine, at the point of execution. The Community Edition was built as a trust machine: open-source signatures, on-device processing, and a browser-based ML demo that let developers test detection quality before asking for anything in return.
That foundation gave us the credibility and the detection engine to build what the enterprise needed next: a complete platform that secures AI traffic (Gateway), AI execution (Sensor), and stays ahead of threats (Intelligence)—all without requiring customer data to leave their infrastructure.
Local-First
Privacy is architecture, not policy. Your prompts, your data, your infrastructure. Zero data transit to third-party clouds.
Research-Powered
Original threat research feeds detection content. Every advisory, every signature, every MITRE ATLAS mapping comes from our own researchers.
Independent
Not acquired, not a feature of a cloud provider, not pressured into data monetisation. Independent security research and product development.