Speaker 1: 0:54

Hello and welcome to. Seo is not that hard. I'm your host, ed Dawson, the founder of the SEO intelligence platform, keywordpupilusercom, where we help you discover the questions people ask online and then how to optimize your content for traffic and authority. I've been in SEO and online marketing for over 20 years and I'm here to share the wealth of knowledge, hints and tips I've amassed over that time. Hello and welcome back to SEO. Iss Not that Hard. It's me here, ed Dawson, hosting, as always, and today I'm going to be talking about model context protocol, mcp servers.

Speaker 1: 1:27

Now, this is a buzzword you might have heard, but it might not be something that you're completely familiar with. What it is, because it is literally so new. So what is it? Well, it's there to solve a particular problem that large language models like Claude Chachi, petit and all those others have in terms of that. They're really really good at generating human-like text, at doing summarization, at doing analysis, that kind of thing, but what they're not very good of is actually doing something outside of writing. So, for example, if you ask it to send an email, fetch data from a database or interact with another application, it won't be able to okay, and lnm by itself is kind of like a closed little box and it can't interact with anything outside of it.

Speaker 1: 2:15

Now, mcp is the technology that comes in that helps bridge the gap between a large language model and the outside world in effect. So other applications, databases, tools, that kind of thing. So what MCP does is it creates a kind of a unified layer, like a bridge, and it bridges the gap between the ai and these external systems, allowing the llm to do more than just brick tech. So instead now it can perform tasks, access databases, interact with various services, without needing an engineer to rework everything for each new tool. So, in other words, it takes away a lot of the headaches involved in connecting an LLM to an external resource and it creates a seamless, standardized way for LLMs to talk to other systems and to understand what other systems can do for it. So yeah, at a basic level, all that an MCP server does is act as that bridge between your large language model and a third-party tool. So, and I should also say that MCP stands for model context, protocol. That's all it stands for. It's just that's what it does. It describes a model, the context and the protocol for how the LLM can talk to the third-party tool in a way that any LLM can use it, so you don't have to rewrite anything for different LLMs, different agents to use this. They just have to plug into them. It's a bit like an API that programmers would use for connecting third-party systems, like a REST API, but it's specifically designed for LLMs to work with, and Anthropic who developed Claude are the people that came up with it and they've made it an open protocol, and OpenAI have said they're going to use it and there are other systems that are also working to integrate with it. So it's looking like it's becoming the new protocol of a standard for LLMs to be able to talk to other tools, and there's loads and loads of people creating MCP servers for their particular tools, their particular services, to be able to work with large language models.

Speaker 1: 4:27

Now We've done it at Keywords People Use. We've created our own MCP server so that Keywords People Use users can connect their Keywords People Use account to their LLM, which then means they can essentially not have to go to the Keywords People Use, do research to get that data and then hand that data to the LLM. You can just do it all directly from the llm. So it means you can say, for example, be in claude and say get me, all the people also ask questions for this particular keyword. So we say so. We say to all the people also ask questions for broadband in the uk. And then claude, using our mcp server, can connect to keywords people use. Do that query, which will then go to scrape that data from Google and bring it back directly into the ILM. You can use this for Google Autocomplete, for semantic keywords, for forum searches, all the kinds of searches that we have in keywords people use can now be natively embedded within your large language model, your large language model, so you can actually allow your chats to interact directly with those services. Now this is great because obviously you can specifically say please get this people source data, please get this google to complete data. But you can also use it with much more powerful prompts, such as saying please create a piece of content, say a broadband beginner's guide, and I'd like you to get Google autocomplete data and people also ask data. But you don't actually have to specify the exact terms to use, because you've told Claude that that's what you want to do and Claude will then decide to do its own research and it can make multiple calls to keywords people use via the MCP server to decide what research to do. So you are essentially now allowing live data from keywords people use to be brought in on demand by the large language model itself, which is really, really powerful and just brings a whole new scope to the kind of things that you can achieve with this.

Speaker 1: 6:45

Now there are lots of mcp server implementations available for all sorts of services. Now, obviously, I've talked about the keywords. People use one, but there are lots of others. There are are ones from places like Google Maps, google Drive, slack Stripe all sorts of places you know, and this is something that's new and emerging really fast Now. Now this allows you to, you know, plug into your LLMs and do more than just create text and analyze text. You can now actually do so much more work with your llm and bring so much more data in directly, and then you know work with external systems and tools as well, which just boosts the whole pile of things you can do with ai now.

Speaker 1: 7:22

Now there are two types of servers. A lot of them actually run locally. So most people talk about mcp servers, they mean ones that run locally. Um, and yeah, that's because the very first ones of these, there was no kind of external server system. You had to literally download them from like GitHub and run them locally. Now, if you've got a bit of technical knowledge it's not completely difficult to do, but if you've never done that kind of thing before, it's quite a learning curve for some people to get there.

Speaker 1: 7:51

So this is the big problem with it and it also has a fault in it in that because you're downloading a local version of a server, if whoever's providing that MCP server, who created the code for it, if they do a bug fix, add some new functionality, anything like that, it's not going to automatically propagate and come to your local version. You have to keep re-downloading as new versions come, which is like imagine having a website and every time you access the website you have to download the whole website to your local computer and you browsed it locally. And if you wanted to get updates if a page was updated computer and you browsed it locally. And if you wanted to get updates if a page was updated, you'd have to go, you know, re-download the entire website again and go and check to do that manually. That wasn't. You know it didn't automatically update. You know you expect to go to website and you know it's updated as an external remote server and is kept up to date doesn't work like that with lots of mcp servers.

Speaker 1: 8:43

There are now, though, mcp servers that are working on this remote basis, so you don't have to download the entire service um to your and run it locally on your computer. You can actually connect to it remotely, and that's how we've done it with our keywords. People use one, so it means, if we bug fix it as we add new functionality to it, you don't have to re-download. It's always going to be the latest up-to-date version. But just be aware that if you start getting into looking at all different kinds of MCP servers, you will find that most of at the moment you have to download and run locally. It's quite rare to find a remote one like the keywords. People use one, but I think is, but soon. I think that all will all be remote before long, because it's just such the more sensible way to do it. You know the more efficient way to do it, um. I think just there's so many people rushing to get them out that the local version was the first first one they did. But just just something to watch out for there.

Speaker 1: 9:37

Okay, now, if you'd like to try this. Try an mcp server, try the keywords people use mcp server. It's really quite simple. All you need to do is go to keywords. People use dot com slash mcp hyphen server. That's keywords. People use dot com slash mcp hyphen server and I'll also put that link in the show notes and there's full instructions there on how you can quickly and easily get the setup and try it out for yourself. You don't need a paid account keywords people use to try it. We've made this open access to people on free accounts as well, so you can try this at no cost. Obviously, on free accounts there's a limited number of credits to use, but you can certainly enough there to get up, test it and see how it works and you know, strongly recommend you do that because it is. It is an actually amazing extension to the functionality that you can get on your LLM. Just to get this working with it is really, really interesting. So do try that.

Speaker 1: 10:32

Yeah, and I'm sure I'll be talking more about MCP servers with QSP and in general. I think this is something that, if it's the first time you've heard of it, I think you're going to hear about it a lot more in the future. Anyway, that's, it's the first time you've heard of it. I think you're gonna hear about it a lot more in the future. Anyway, that's it for today. So you know, as always, remember, keep optimizing, stay curious and remember seo is not that hard when you understand the basics. Thanks for listening. It means a lot to me.

Speaker 1: 10:58

This is where I get to remind you where you can connect with me and my seo tools and services. You can find links to all the links I mentioned here in the show notes. Just remember, with all these places where I use my name, that Ed is spelled with two Ds. You can find me on LinkedIn and Blue Sky. Just search for Ed Dawson.

Speaker 1: 11:12

On both you can record a voice question to get answered on the podcast. The link is in the show notes. You can try our SEO intelligence platform, keywords People Use at keywordspeopleusecom, where we can help you discover the questions and keywords people are asking online, pus those questions and keywords into related groups so you know what content you need to build topical authority. And, finally, connect your Google Search Console account for your sites so we can crawl and understand your actual content, find what keywords you rank for and then help you optimise and continually refine your content and targeted, personalised advice to keep your traffic growing. If you're interested in learning more about me personally or looking for dedicated consulting advice, then visit wwweddawsoncom. Bye for now and see you in the next episode of SEO. Is not that hard.