VicOne xPhinx is an edge AI security solution designed specifically for in-vehicle AI systems and smart cockpits. The product protects edge AI models and AI agents from prompt injection attacks, jailbreak attempts, unsafe behaviors, and data leakage without requiring modifications to existing AI models. xPhinx operates directly on the vehicle device with 100% local processing, eliminating the need for cloud connectivity. The solution uses a dual-layer, risk-aware architecture where a lightweight first layer runs continuously, while deeper intent analysis activates only when higher-risk behavior is detected. This approach delivers up to 70% faster execution and up to 90% lower memory usage compared to LLM-based guardrails. The product inspects and sanitizes LLM and VLM inputs and outputs to prevent manipulated or unsafe behavior at the point where AI decisions are made. It supports multiple hooking methodologies to intercept LLM inputs and outputs across different AI frameworks and operating systems. xPhinx is powered by automotive threat intelligence that keeps pace with evolving prompt attacks and jailbreak techniques. The solution is developed under ASPICE CL2 processes and supports risk management aligned with ISO/SAE 21434 and UN R155 automotive cybersecurity standards. It operates entirely offline and maintains data residency within the vehicle.