The Cloudflare API MCP in Under 3 Minutes

We connect an AI coding agent to the entire Cloudflare API, buy a domain, deploy an application, diagnose a production error, fix it, redeploy, and verify the result—all in under 3 minutes. In this video, I set up Cloudflare’s API MCP server and give my coding agent access to my Cloudflare account using OAuth. The agent finds an available domain, purchases it, deploys a React Worker, checks Workers Observability for production exceptions, fixes the error, and redeploys the application—all without me writing any code. The Cloudflare API MCP gives AI agents access to more than 2,500 API endpoints across Workers, DNS, R2, D1, Zero Trust, and the rest of Cloudflare’s platform. It works with MCP-compatible tools and agents, including Claude Code, Codex, Pi, and more. Only grant your agent the Cloudflare permissions you trust. I use full access for this demo, but you can select more restricted permissions during authorization. Create a Cloudflare account for free: https://dash.cloudflare.com/sign-up Tools and resources mentioned: The site: https://mcpboom.win: The repostiroy: https://github.com/jillesme/cloudflare-mcp-code-mode-demo Cloudflare API MCP GitHub repository: https://github.com/cloudflare/mcp Cloudflare Workers: https://developers.cloudflare.com/workers/ Workers Observability: https://developers.cloudflare.com/workers/observability/ Cloudflare Registrar: https://developers.cloudflare.com/registrar/ Cloudflare API documentation: https://developers.cloudflare.com/api/ Chapters: 00:00 Intro & triggering the error 00:21 Installing the Cloudflare API MCP 00:35 Authenticating with Cloudflare 00:55 Finding an available domain 01:16 Buying the domain & deploying 01:31 Triggering the production error 01:35 Checking Workers Observability 01:48 Fixing the error & redeploying 01:54 Testing the fix 02:02 Verifying the production logs 02:08 What we accomplished 02:22 What the Cloudflare API MCP can do 02:52 Conclusion #cloudflare #mcp #cloudflareworkers #aiagents #developer