Mangools now has an MCP server: SEO data straight in ChatGPT, Claude and Cursor

Mangools has launched an MCP server. It lets you connect ChatGPT, Claude, Cursor and other AI assistants directly to your Mangools account, so you can pull keyword research, SERP analysis, rankings, backlinks and AI visibility just by asking for it. No export to a spreadsheet, no copying and pasting between tabs.

We use Mangools ourselves every day for research, so this integration was immediately relevant to us. Below is what it actually is, what it gets you, and where it still doesn't replace your own judgement.

What is an MCP server, exactly?

MCP stands for Model Context Protocol: a standard way to give an AI assistant access to an external source. On its own, your AI assistant can't look anything up beyond what it knows or finds on the internet. With an MCP connection, it can call a real tool and use the outcome in its answer.

The difference is significant. Without the connection, an assistant guesses at search volumes. With the connection, it pulls the real volume, the difficulty score and the current top 10 from Mangools and works from that.

What the Mangools MCP server can do

The server gives access to 78 endpoints, spread across all six tools in the subscription:

• KWFinder — keywords, search volume, difficulty and CPC
• SERPChecker — who currently ranks positions 1 to 10, and why
• SERPWatcher — your own positions and how they're moving
• LinkMiner — backlinks for you and your competitors
• SiteProfiler — domain overview and authority
• AI Search Watcher — whether and how your brand shows up in AI answers

In practice you just type a question: which keywords around this topic are achievable, which of my keywords dropped this week, what are the strongest links to this competitor, set up AI monitoring for my brand. Your assistant picks the right endpoint itself, retrieves the data and delivers an answer with the numbers underneath.

Connecting it: logging in instead of copying keys

It's a remote server, so there's nothing to install. In every client it's the same three steps: paste the server URL, log in with your Mangools account, and approve access. No API token to hunt down, no config file to edit.

That login works per user and on every plan. What's new is that this now also works in ChatGPT: its own MCP connectors only accept this login method. Step-by-step instructions for Claude, ChatGPT, Cursor, Microsoft Copilot Studio, n8n and Open WebUI are on Mangools' MCP page.

Your API token isn't going anywhere. It remains the way to use the REST API and is the alternative for clients that don't support the login method.

And it got five times cheaper

An API or MCP request now costs 2 in-app lookups instead of 10. That's five times as much research through your AI assistant within the same subscription. There's nothing extra to buy: requests come off the daily limit already included in your plan, whether you log in or use a token. No separate credits, no add-ons.

What we actually use it for

Where this genuinely saves time:

1. Sizing up a topic quickly. Volume, difficulty and the current top 10 in one question, without first clicking through four screens.
2. A weekly check on drops. Ask which tracked keywords fell, and have your assistant immediately add which page belongs to each.
3. Building clusters. Grouping raw keyword lists by intent is exactly the kind of work a language model is good at — now based on real numbers instead of a guess.
4. Comparing link opportunities. Who links to three competitors and not to you: that's a list you can act on straight away.

And where you need to watch out:

• An assistant jumps to conclusions too fast. The data is real, the interpretation stays your job. Especially with low volumes and with AI visibility, where the numbers vary from measurement to measurement.
• Check which period and location it's using. "Hairdresser Eindhoven" viewed from Amsterdam gives a different picture. Name the location and period explicitly in your question.
• Your limit can be burned through. An assistant that calls thirty endpoints back-to-back to answer a single question uses up your daily limit faster than you'd think. Ask for something specific.

Who is this useful for?

If you already work with Mangools and use an AI assistant every day, this is a half hour of setup that pays for itself. If you're not yet using an SEO tool, this is a good reason to start: Mangools is the most logical first step in terms of price and ease of use, and you can do your first bit of research straight away in KWFinder.

Want to know how this tool compares to the bigger packages, read SEO tools compared: Semrush, Ahrefs and Mangools. More on visibility in AI answers is in is SEO dead or is it changing.

Frequently asked questions

What is MCP?

MCP (Model Context Protocol) is a standard that lets an AI assistant call an external tool and use the outcome in its answer. Without such a connection, the assistant can only work with what it already knows.

Do I need an API token for the Mangools MCP server?

No. You paste the server URL into your AI client, log in with your Mangools account, and approve access. A token is still needed for the REST API and for clients that don't support logging in.

Does it work on every Mangools plan?

Yes, logging in works per user and on every plan. Requests use the daily limit included in your subscription; an MCP request costs 2 lookups instead of 10.

Does it also work in ChatGPT?

Yes. That's new: ChatGPT's own MCP connectors only accept the login method, and Mangools now supports it. Claude, Cursor, Microsoft Copilot Studio, n8n and Open WebUI also work.

Does this replace an SEO specialist?

No. It retrieves data faster and saves clicking around. Deciding which pages to create, what to prioritise and whether an outcome is meaningful stays human work, and that's exactly where the difference in results lies.

Want to think through how to set up this kind of tooling without ending up with a pile of separate subscriptions? Plan a free call, and we'll take a look with you.