Cloudflare AI

Interactive visual walkthroughs of Cloudflare’s AI platform. Start with the reference architecture, or jump into a category.

Featured Reference Architecture

How Cloudflare ships code with AI — on its own platform

Cloudflare’s public blog posts describe AI Code Review across standard CI pipelines — with the full internal AI engineering stack underneath: Access, Proxy Worker, AI Gateway, MCP Server Portal, Backstage, Engineering Codex, Workers AI, and seven specialised review agents. Walk through the same programmable primitives customers can compose into their own secure AI development workflows.

Public blog stats As of April 2026 · 30-day windows

3,683 active users (93% R&D)

$1.19 avg cost / review

85.7% cached-token rate

+58% MRs / week

241B tokens / month

Source: public Cloudflare Blog posts “The AI engineering stack we built internally — on the platform we ship” and “Orchestrating AI Code Review at scale.”

Cloudflare Access Workers AI Gateway Workers AI MCP Portal Code Mode Agents SDK Logpush OpenCode

Explore the reference architecture →

Or explore by category

🛡

7 Use Cases

AI Security

Learn how Cloudflare secures AI workloads — governing employee GenAI usage, protecting AI-powered applications, securing self-hosted agents, and orchestrating safe multi-agent communication.

Gateway Zero Trust WAF DLP CASB AI Gateway Agents SDK

Explore use cases →

⚡

9 Use Cases

AI Builder

Learn how to build AI applications on Cloudflare — API key management, dynamic routing, RAG, voice agents, persistent chat, scheduled agents, web-browsing agents, private networking (Mesh), and durable long-running agents (Project Think).

AI Gateway Agents SDK AI Search Workers AI Browser Run Cloudflare Mesh Project Think

Explore use cases →