Cloud.Red MCP
An MCP for your F5 BIG-IP fleet.
Stop hunting through portals. SSH into any device, push a config, or open a web console with a single command. SSO-secured. AI + MCP ready. Installs in under 2 minutes. Enabled by Cloud.Red CLI.
Your AI Assistant.
Your F5 Estate.
Finally Connected.
Cloud.Red MCP gives any AI assistant secure, real-time access to your F5 and NGINX infrastructure — so you can ask questions in plain English and get answers in seconds.
Six clicks become one line.
Without crcli
With crcli
Stop doing this the hard way.
crcli replaces fragmented, manual workflows with a single unified tool.
Three reasons your team will
never go back to the portal.
That pool health check script already lives in your repo, in version control, next to the rest of your code. Cloud.Red MCP means you don't open a console to run it. You just ask. Claude reaches into your fleet through the MCP connector, finds both BIG-IPs by hostname (no UUIDs), runs your script against each one, and reads the results back to you — pools with no monitor, every member that's down, and the exact reason each health check failed. Across both devices. In one ask. Every action logged.
- "Run my pool health check against both BIG-IPs in prod-ha and summarize"
- "Which pools have no monitor configured across my fleet?"
- "List every pool member that's down and why the health check failed"
Chasing a connection problem used to mean SSH-ing to the box, remembering the exact tcpdump flags, and squinting at a wall of packets. With Cloud.Red MCP, you describe what you want to see. Claude runs the capture live on the BIG-IP, 50 packets to that host on that port, using the correct F5 interface syntax. Then it reads the capture back: who's talking to whom, the handshake, the retransmit, and exactly where the conversation stalls. Not raw output. An explanation of what it means. Without leaving the conversation.
- "Capture 50 packets to 10.1.2.50 on port 443 on bigip-prod-01"
- "Read that capture and tell me where the connection is stalling"
- "Summarize the handshake and flag any retransmits"
Point Cloud.Red MCP at your tenant with one command and it knows your fleet: your devices, your groups, your permissions. Then log in once. Browser SSO completes and that single session covers everything, the CLI, your scripts, and AI through the MCP connector. No copying configs between windows, no portal tab, no separate credentials. Hostnames, not UUIDs. And whether you typed the command or AI did, every action is logged under the same authenticated identity.
- crcli setup <your-tenant> ← one command, fleet discovered
- crcli auth login ← browser SSO, covers CLI + scripts + AI
- crcli devicegroups ssh prod-ha -S ./pool_health_check.py ← both BIG-IPs, one line
Everything your team needs.
One binary. Zero dependencies.
Download the binary for your platform and drop it in your PATH.
Up and running in under 2 minutes.
Download the Binary
One static binary. No runtime. No package manager required.
Point It at Your Tenant
Your tenant name is the subdomain of your Cloud Red portal URL.
$ crcli setup <your-tenant>Authenticate via SSO
Browser opens. Complete your login. Token stored in keychain. Done.
$ crcli auth login20 minutes. That’s all we need.
See Cloud.Red MCP running against a live fleet. No slides. No pitch deck. Just a terminal and your questions.