Figuring out WP Engine’s Good Seek AI Style Context Protocol (MCP) Server

The Good Seek AI MCP Server is a formidable new function in WP Engine’s AI Toolkit that transforms your WordPress®¹ website right into a dynamic, real-time wisdom base for any exterior Huge Language Style (LLM) you connect with it. When enabled, this server responds to requests from AI gear formatted the use of the Style Context Protocol (MCP) usual.

On this article, I can quilt what MCP is, how you can paintings with the Good Seek AI MCP Server, and the way it complements the Good Seek AI plugin, which is to be had for all WordPress websites.

What’s MCP and the way does it paintings?

Style Context Protocol (MCP) is a standardized conversation protocol that connects AI fashions to are living, exterior information assets and gear. Whilst massive language fashions like ChatGPT are extremely clever, their wisdom is restricted to their coaching information and is frozen in time, which means they may be able to’t get entry to real-time or explicit area of interest information.

On its own, it most probably does no longer learn about your corporate’s newest product specs, the present inventory value, or the content material of a weblog put up you revealed this morning. MCP is the bridge that closes this hole between the AI’s static wisdom and the dynamic, real-time international.

To grasp its function, recall to mind MCP because the USB (Common Serial Bus) for synthetic intelligence. Sooner than USB, connecting a printer, mouse, or keyboard to a pc required a complicated array of various ports and customized drivers. MCP addresses a equivalent drawback within the AI ecosystem. With out a usual, connecting an AI to each and every unmarried web site, database, or interior API will require writing customized, one-off integrations—a posh and inefficient procedure.

MCP supplies that common usual. It defines a collection of straightforward, predictable regulations and instructions, permitting any AI style to seamlessly “plug into” any MCP-compliant information supply. The AI doesn’t want to know the complicated inside workings of your web site’s database; it simply must understand how to “discuss MCP.”

Thru this protocol, the AI can successfully factor standardized requests like “Seek your wisdom base for this time period” or “Fetch the contents of this explicit web page”. In essence, MCP transforms a static, encyclopedic AI right into a dynamic, context-aware agent in a position to getting access to and reasoning about your present content material.

Good Seek AI MCP Server

The Good Seek AI MCP Server, which is enabled through default and calls for person opt-out, exposes “fetch” and “seek” gear that permit exterior AI fashions to engage with a web site’s public, revealed content material.

Why will have to you utilize MCP?

The principle advantage of the use of MCP is that it permits builders to glue exterior AI brokers, like ChatGPT or Claude, at once to their web site’s are living content material. This empowers them to construct complicated AI programs, corresponding to customized chatbots or assistants, which can be grounded in real-time, correct data from their very own website as a substitute of the generic, frequently old-fashioned information the fashions have been educated on.

The way it works with WP Engine’s AI Toolkit

The Good Seek AI MCP Server is a function of the Good Seek AI carrier. Whilst you allow it, this server listens for requests which can be formatted the use of the MCP usual. Right here’s a regular workflow:

A query is requested: A person interacts with an AI utility, like a customized chatbot constructed with Claude or ChatGPT.
The AI wishes extra data: The AI style realizes it wishes present data to respond to the query correctly. It sees that it has get entry to to an MCP server that gives gear for getting access to your web site’s information.
The AI style sends a community request on your web site’s distinctive MCP server deal with. For instance, it could ask your server to seek for “details about AI Toolkit.”
The MCP server receives this request, makes use of the Good Seek AI vector database to search out essentially the most related content material in your web site, after which sends that data again to the AI style.
The AI style now has contemporary, correct content material out of your website. It makes use of this data to formulate a related, up-to-date solution for the person.

Briefly, the MCP server permits AI programs to be powered through the real-time, correct data out of your web site’s semantic seek and vector database, fairly than the stale or scraped information that an LLM would differently be restricted to. This turns your web site right into a are living, dynamic wisdom base for any AI agent you connect with it.

Checking out the Good Seek AI MCP

After getting an MCP server operating, the next move is to connect with it and check its features. That is the place a shopper inspector turns into helpful. Gear like the MCP Inspector or a flexible API consumer corresponding to Postman (the use of its WebSocket request function) will let you have interaction along with your server simply as an AI style would. This procedure is necessary for debugging and making sure your server supplies the right kind information.

On this article, I’m stoked about the use of Postman as a result of this is a bit more straightforward to paintings with, I feel.

Step 1: Organising a connection

First, you wish to have the original URL to your Good Seek AI MCP Server (consult with the directions on how you can download it right here). In Postman, you might click on at the “New” button on the best of your Workspaces merchandise web page. This may display a card menu. Click on on “MCP” to create the interface web page to engage along with your server. It looks as if this:

It’s going to take you to the MCP interface web page. That is the place you’ll paste your URL within the deal with bar, then click on “Attach.”

A a hit connection is indicated through a standing message. You’ll see the golf green “Attached” notification on the backside of the Postman window. This confirms that your consumer is now actively paying attention to the MCP server.

Step 2: Finding the to be had gear

That is the place the navigation starts. As soon as hooked up, the MCP server right away advertises the gear it makes to be had. Those gear are the particular purposes or movements the AI is authorized to accomplish. Recall to mind them as API endpoints, however for an AI.

Within the screenshot, we will be able to see the server has offered two distinct gear:

fetch: The outline says, “Fetch a particular put up through its ID from the Elasticsearch index.” It is a extremely explicit device that calls for a novel identifier to retrieve a unmarried piece of content material.
seek: The outline is, “Seek for data within the hooked up Elasticsearch index, please attempt to refine the hunt question up to conceivable.” It is a extra versatile device designed for querying the knowledge supply with herbal language or key phrases.

This discovery segment is prime to MCP. The customer (and through extension, the AI) doesn’t want prior wisdom of what the server can do; the server broadcasts its personal features.

Step 3: Creating a request (Interacting with a device)

Now that we all know what gear are to be had, we will be able to ship a message to make use of certainly one of them. MCP messages are normally formatted in JSON, specifying the tool_name to make use of and the arguments it calls for.

Let’s check the seek device. For this situation, I can simply use the filter out: string enter, which accepts freeform textual content. I typed “webinar” within the enter field as a result of my WordPress content material accommodates a webinar put up. At the aspect JSON pane, it looks as if this:

This JSON object explicitly tells the MCP server: “Use your seek device and provides it a question with the price “webinar.“

Step 4: Figuring out the reaction

After you ship the request, the MCP server will execute the device with the arguments you supplied and ship a reaction again. This reaction is the uncooked information that the AI would obtain to formulate its solution.

That is the a hit reaction we get again:

{

 "content material": [

   {

     "type": "text",

     "text": {

       "results": [

         {

           "id": "doc-1",

           "title": "Webinar – WP Engine MCP",

           "text": "This is WP Engine's Webinar Show about nerd stuff",

           "url": "https://demo.example.com/webinar/getting-started"

         }

       ]

     }

   }

 ]

}

The form that comes again is a stringified JSON blob with a effects array. I parsed it and put it in a code block to make it extra readable for this newsletter.

The Good Seek AI MCP reaction is an object with a unmarried content material array, the place each and every part represents one piece of output.

On this instance, a content material merchandise has kind: “textual content” and a textual content payload this is an object containing a effects array. Each and every access in effects is a record with 4 core fields: identification , name, textual content and url .

This envelope makes it simple to movement or mix more than one output portions, whilst the internal effects items will also be prolonged (e.g., upload rating, website, or published_at) with out converting the outer form.

Attach Good Seek AI MCP at once into your AI style

You’ll disclose Good Seek AI to any AI assistant through operating it as an MCP server and plugging it in via a connector.

I can display Claude on this instance. We want to sign up the similar endpoint the use of their MCP consumer configuration.

As soon as hooked up, the style can uncover Good Seek AI’s declared gear by way of MCP’s tool-listing protocol and make contact with them with structured arguments—no bespoke SDK required.

Functionally, wiring Good Seek AI via MCP upgrades your assistant from “best-effort guessing” to retrieval-augmented solutions which can be exact, auditable, and policy-aware.

When you use Claude, the upload a customized connector web page looks as if this:

As soon as hooked up, your AI style will know the skills and tooling it could actually name on out of your website. It’s going to be uncovered within the dropdown selector.

Your website seek, reimagined

The adventure from a static webpage to an interactive, clever useful resource is the following nice soar in virtual reports. We’ve observed how the MCP acts as a connector, bridging the space between AI and the real-time, precious content material in your web site. When this protocol is blended with the semantic energy of WP Engine’s Good Seek AI, your WordPress website is now not only a vacation spot for customers; it turns into a dynamic information supply that any AI agent can seek the advice of.

By means of offering the gear to combine your content material with the arena’s maximum complicated AI fashions, WP Engine is hanging you at the vanguard of this new AI age. Permit the Good Seek AI MCP server on any WordPress website and get the facility of AI and WordPress!

WP Engine is a proud member and supporter of the group of WordPress® customers. The WordPress® emblems are the highbrow assets of the WordPress Basis, and the Woo® and WooCommerce® emblems are the highbrow assets of WooCommerce, Inc. Makes use of of the WordPress®, Woo®, and WooCommerce® names on this web site are for identity functions most effective and don’t indicate an endorsement through WordPress Basis or WooCommerce, Inc. WP Engine isn’t counseled or owned through, or affiliated with, the WordPress Basis or WooCommerce, Inc. ︎