WP Rocket Adds an MCP Server: Now Claude and Codex Can Read Your Site's Performance Data
WP Rocket is becoming the first WordPress performance plugin with an MCP server — letting AI assistants like Claude and Codex read Rocket Insights scores and adjust WP Rocket settings, with your approval. Here's what it does and why it matters for agencies.

Table of contents
WP Rocket is adding something no other WordPress performance plugin has: an MCP server. It lets AI assistants such as Claude and Codex connect directly to a site's real performance data and to WP Rocket's own settings — read scores, review recommendations, and propose configuration changes you approve before anything is applied. It ships at no extra cost in every WP Rocket license.
If you manage more than a handful of WordPress sites, that sentence is worth reading twice. The tedious part of performance work — checking each site, comparing scores, applying the same fixes across a portfolio — is exactly the part an AI assistant with direct access can compress from hours into minutes.
What is MCP, quickly
MCP — the Model Context Protocol — is an open standard for connecting AI assistants to real tools and data instead of leaving them to guess. Without it, an assistant can only act on what you paste into a chat. With an MCP server, the assistant can query live data and call defined actions directly. WP Rocket MCP is exactly that: a bridge between your AI assistant and the plugin, so the assistant works from your site's actual numbers rather than a screenshot.
Crucially, the connection is read-and-propose, not read-and-execute. Any change the AI suggests to a WP Rocket setting waits for your approval before it's applied. You stay in control at every step.
What you can actually do with it
WP Rocket frames the launch around jobs that agencies and freelancers do repeatedly:
- Monitor performance across many client sites from one AI conversation instead of logging into each dashboard.
- Pull Rocket Insights scores, reports, and recommendations without hopping between tabs.
- Review and adjust WP Rocket settings — the assistant proposes, you approve.
- Roll out a proven configuration across similar sites, so a setup that works on one client's WooCommerce store can be replicated on the next.
- Generate client-ready, branded performance reports on demand.
- Spot optimization opportunities and measure the result automatically after a change.
WP Rocket's own estimate is that tasks which used to take 60–90 minutes per site can drop to 5–10 minutes through a single AI conversation. Treat that as a vendor figure, not a guarantee — the real saving depends on how uniform your sites are — but the direction is right: the more sites you run, the more the per-site overhead is what actually eats your week.
Why this lands now
There's a strategic read here that's worth naming. As AI makes building a website faster and cheaper, the durable value for agencies and freelancers shifts from launch to everything after it — maintenance, optimization, and the monthly report that justifies a retainer. A tool that makes recurring performance work fast and repeatable is a tool that makes that retainer easier to defend.
WP Rocket MCP leans directly into that. It turns "check and tune 30 client sites" from a dreaded afternoon into a chat, and it produces the branded report at the end. For a maintenance-plan business, that's less a feature and more a margin lever — and it's an industry-first, so it's a genuine point of differentiation while competitors catch up.
The honest caveats
A few things to keep in mind before you build a workflow on it:
- You still own the decisions. The approval gate is the point — don't rubber-stamp a batch of AI-proposed changes across 30 sites without understanding what each one does. A caching or delay-JS change that's perfect for a brochure site can break a checkout.
- It reads WP Rocket's data, not the whole web-vitals picture. Rocket Insights is a strong signal, but real-user metrics (field data) and a proper before/after check still matter for anything you're billing on.
- It's brand new. Model access to live settings is powerful; give it the same care you'd give any tool that can change production. Test the flow on a staging site before you point it at a client's live store.
Used with that discipline, though, WP Rocket MCP is one of the clearest examples yet of AI doing the unglamorous, repetitive part of the job — and leaving the judgement to you. If performance maintenance across a portfolio is how you bill, it's worth setting up early.
For the current WP Rocket features that MCP surfaces — the built-in CDN, caching, and file optimization — see our look at what RocketCDN adds inside the plugin.
Sources
- WP Rocket Affiliate Program — WP Rocket MCP announcement (sneak peek) wp-rocket.me
- Model Context Protocol — open standard for connecting AI assistants to tools and data modelcontextprotocol.io
- WP Rocket — Rocket Insights and plugin features wp-rocket.me


