Hey!!! We're the maker of SentinelAI. The idea came from a frustrating realization while building AI applications: we obsess over model quality before launch, but once AI reaches production, we often have very little visibility into what it's actually doing. A model can hallucinate, leak sensitive information, get jailbroken, or behave unexpectedly, and many teams only discover it after users report it. That felt backwards. We monitor servers, databases, APIs, and infrastructure. Why not AI behavior itself? So I started building SentinelAI: a platform that analyzes prompts and responses, detects risky patterns, surfaces anomalies, and helps teams understand when their AI systems may be drifting into unsafe or unreliable territory. What started as a side project quickly turned into a bigger question: How do we build AI systems that are not just powerful, but observable and trustworthy? That's the problem I'm exploring with SentinelAI. I'd love to hear your feedback, and if you're building with LLMs or AI agents, what's been your biggest challenge in monitoring them once they're go live? Thanks for checking it out!

Sentinel AI

Spot Risky AI Behavior Before It Ships

Spot Risky AI Behavior Before It Ships