Integrity × Cloudflare — AI Security & Capability Showcase

cloudflare-integrity.com · live demos served entirely from Cloudflare's edge Enterprise no origin server

Your request, as Cloudflare sees it

This page is a Cloudflare Worker. Every value below was computed at the data center nearest you - before any server was involved.
Loading...
-

AI Security Playbook

live controls from the 2026 Enterprise AI Security Playbook
Secure AI Gateway Live
Input/output guardrails in front of a real LLM: prompt-injection blocked, PII redacted before egress, keys server-side, rate-limited.
Ch.6 - AI Gateway + Firewall for AI
DLP for GenAI Live
Inline inspection of a prompt before it leaves - six data classes plus intent, with block / redact / allow.
Ch.8 - Data protection
Agentic AI & MCP Live
Tool poisoning, injection-via-tools, and excessive agency - caught by the MCP portal, least-privilege, and a per-invocation audit log.
Ch.7 - Securing agents & MCP
Shadow AI governance Live
Discover, score (1-5), and tier every AI tool in use - sanction, monitor, or block - with a user coaching page.
Ch.4/5 - Discovery & governance
OWASP LLM Top 10 & maturity Live
The board-briefing reference: 10 LLM risks mapped to controls, plus the L1-L5 maturity model.
Ch.3/10 - Taxonomy & maturity

Platform capabilities

core Cloudflare, live
AI at the edge Live
A chatbot powered by Workers AI - an LLM runs in Cloudflare's network, no backend, no GPU to manage.
Security playground Live
SQLi / XSS blocked at the edge, live bot-score gating, and rate limiting in action.
Zero Trust Access Live
An internal app fronted by a Cloudflare Access identity wall - one-time PIN login before anything loads.
Edge personalization Live
Geolocation greeting, A/B split, and device detection - decided at the edge in microseconds.
Load Balancing / Tunnel Planned
Failover and no-public-IP access - pending an extra zone-level token permission.