AI Connector in Blog Tools – Blog Settings with Claude and ChatGPT

Blog Tools AI connector as a control panel
Summarize with AI:

BeeClear is a maker of WordPress plugins and apps, and an SEO and GEO agency. Blog Tools is a BeeClear plugin for WordPress that brings a blog’s content and SEO layer into one place: table of contents, author box, meta titles, structured data, Open Graph, redirects and a tag manager. From version 2.0 the plugin has an AI connector. The AI connector is a link through which Claude, ChatGPT, Codex or Gemini read and change Blog Tools settings in an ordinary conversation. Instead of hunting for options across panel screens, you type a request and the model checks the configuration and makes the change. Below we explain how the connector works, what it covers and how to switch it on safely.

What is the AI connector in Blog Tools?

The AI connector in Blog Tools is an MCP server built into the plugin. MCP is the Model Context Protocol – an open standard for connecting AI assistants to external tools. The connector hands the model 30 described tools, and the model calls them on your behalf. Each tool matches one action from the panel – for example reading a module’s settings or adding a redirect. The server runs on the WordPress REST API at /wp-json/beeclear-blog-tools/v1/mcp and is off by default.

For you, it mainly saves time. Blog Tools has more than twenty modules, each with its own options. The model knows all of them from the first message, because the connector gives it the current values together with the defaults. You ask about your site the way you would ask a colleague, and you get a concrete answer right away.

What can an AI assistant do in Blog Tools?

The connector gives the model every Blog Tools feature – from module switches to the fields of a single post. A read-only token lets it browse, while a read-and-write token also lets it make changes. Here is what the AI sees and changes in each area:

AreaReadWrite (read-and-write token)
Plugin modulesswitch states, every option with its defaultswitching modules, changing options, importing a configuration
Post SEO and Open Graphtitle, meta description, robots, social cardfilling in and clearing post or page fields
Redirectsrules with hit counts, conflicts, the 404 logadding, changing and removing rules
Structured dataJSON-LD values, sets and rulesadding, changing and removing items
Tag manager and tooltipstag code, placement and loading modeadding and changing tags and tooltips
Live pagetitle, description, canonical, Open Graph, JSON-LD–

The last row of the table is a checking tool. The connector fetches a page of your site the way a visitor does and shows what actually ended up in the code. Claude or ChatGPT uses it to verify the effect of a change without guessing.

How do you connect Claude or ChatGPT to Blog Tools?

You connect Claude or ChatGPT to Blog Tools in a few minutes from the Blog Tools → AI connector screen. The screen shows ready-made commands for each client. The whole setup comes down to five steps:

  1. Switch the connector on with the button on the AI connector screen.
  2. Create a token with its own name and choose read-only or read-and-write access.
  3. Copy the token or the secret connector URL – the plugin shows them only once.
  4. Add the connector in your client: paste the secret URL in claude.ai or ChatGPT, or run the given command in Claude Code, Codex or Gemini CLI.
  5. Ask the chat for a Blog Tools overview and check that it answers.

In claude.ai and Claude Desktop you add it as a custom connector without OAuth sign-in. ChatGPT needs developer mode switched on in its connector settings. Command-line tools take the token in the Authorization header, which is a safer route than the secret URL.

Which requests can you give the AI connector?

You give the AI connector requests in plain language, just as in any other chat. “Show posts without a meta description” runs a post search and a check of their SEO fields. “Redirect /old-offer/ to /offer/” creates a 301 rule, and the plugin refuses it if it would close a loop. In turn, “Check what Google sees on the pricing page” returns that page’s title, description, canonical and JSON-LD blocks.

Requests that combine reading with a decision work best. Ask for a report first, review the suggestions, and only then approve the write. The model sees the previous values, so it can describe exactly what it will change. If something goes wrong, ask it to restore the latest backup – that takes one message.

How does the connector protect your settings from AI mistakes?

The connector protects your settings in three ways: validation, backups and a log. Every value from the model goes through the same rules as the panel form. An invalid default redirect code turns into 301, and an unknown settings key ends with a clear error. Tools that write settings and post fields also have a dry-run mode, in which the model shows the result without saving it.

Before every change of settings, structured data, tags and tooltips, the plugin backs up the configuration. You can restore any of the last twenty backups with one request. The log records every change with the token name, and for redirects and post fields also the previous values. You see the latest entries on the connector screen as well.

Tokens have limits too. The site stores only their hash, so nobody can read a lost token – you create a new one. A token works only while your account still manages the site. Tag manager scripts are saved only for an account allowed to post unfiltered HTML, and the AI cannot touch the “delete data on uninstall” switch at all.

The Blog Tools connector differs from the Smart Link Manager connector in scope. Smart Link Manager is a BeeClear plugin for automatic internal linking, and its connector manages only phrase rules and links. The Blog Tools connector covers the rest of the SEO layer: meta titles, descriptions, structured data, redirects and Open Graph. The two servers have separate addresses and separate tokens.

You can add both connectors to one AI client. Then, in a single conversation, you can fix a post’s meta description, redirect its old address and add a linking rule for a new page. To the model, these are simply two sets of tools with different names.

What do you gain from Blog Tools and the AI connector?

With Blog Tools and the AI connector, blog SEO housekeeping stops being a separate project. You fix missing meta descriptions, redirects and structured data in a conversation instead of clicking through plugin screens. The model also checks what actually made it into the page code. Your decisions rest on the state of the site, not on guesswork.

The AI connector comes with every Blog Tools license, together with redirects, structured data and Open Graph. You can try the full version on your own blog for three days. Your SEO layer can talk to AI from day one.

Summarize with AI:

Want clients and AI models to find your website?

We run website SEO for Google and for AI answers. We tidy structure, internal linking and content, and we measure results instead of promising them.