Introduction
Veriloop is a closed-loop adaptive security layer for your APIs and product. You send it events, it returns a risk decision, and it gets sharper with every label your team applies — without any external AI APIs.
What you get
Two HTTP endpoints — POST /api/public/v1/evaluate and POST /api/public/v1/feedback — plus a dashboard for reviewing events, labelling them, and watching the per-workspace model evolve.
What Veriloop is not
It is not a static rules engine, not a third-party fraud blacklist, and not a hosted LLM. Every signal weight, behavior profile, and trust prior is derived from your workspace's event history and the feedback your team gives.
Where to go next
Quickstart →
Send your first event in 60 seconds.
API reference →
Every field of every endpoint.
Integration recipes →
Express, Next.js, gateway, browser HTML.
How it learns →
The closed loop, in plain English.
Need a key? Sign up and create a workspace — the first key is shown to you exactly once.