I Built an MCP for SEO. Then I Analyzed My Own Entity.

SEO consultant Sara Taher built a Model Context Protocol (MCP) server for Google Knowledge Panel that queries the Google Knowledge Graph Search API and Wikidata to find entities, benchmark competitors, audit a single entity and generate schema. Running the MCP against her own name, "sara taher," Claude found 20 entities sharing the name and identified another Sara Taher as the most prominent, not Taher herself. Taher published the tool and its installation steps on her site and the source code in a GitHub repository, noting users need a Google Knowledge Graph API key.

If you find value in my newsletter and want to support the work that goes into it consider — ☕ Buy Me a Coffee https://buymeacoffee.com/sarataher ☕ 💡 Want more from me? SEO Strategy Course https://sarataher.podia.com/how-to-create-an-seo-strategy | Build Your Own SEO Tools with Python & Agentic AI https://sarataher.podia.com/introduction-of-python-for-marketers | Premium SEO Mastermind https://sarataher.podia.com/in-house-seo-enterprise-meetup Alternatively you can become a paid subscriber: Intro I have a lot of WIP SEO projects, and I think it’s a nice problem to have. Because you can get bored of one, and simply switch to another, or just start another one 😄 One of those projects that have been waiting for a while is building an MCP for SEO. MCPs or Model Context Protocol are easy ways to connect your AI model of choice e.g claude to your favorite SEO tools so that you can have a conversation with the AI while it retrieves and uses real data from those tools to answer your questions. In my case, I decided to build an MCP for Google Knowledge Panel https://www.sara-taher.com/tools/google-knowledge-panel-mcp so buckle up here’s what and how I did it Building an MCP for SEO I wanted to build an MCP for Google trends and somehow pivoted to Google knowledge Panel MCP. First things first, I defined 4 functions I want this MCP to be able to do: - Find every entity behind a name. - Work out which entity is strongest. - Benchmark competitors. - Audit one entity in-depth. - Generate schema for an entity. To achieve this, I needed two things: 1 Google Knowledge Graph Search API: what Google itself knows about an entity. This is the graph behind knowledge panels. 2 Wikidata: the free, open knowledge base Google has long drawn on. It holds how the entity connects to the real world. Google tells you that an entity exists and how Google describes it. Wikidata tells you who it is connected to, what it’s linked to and how well-documented it is. After working with claude code a bit, some testing, etc… the MCP came to life and now it was time to actually use it. I downloaded claude for desktop I usually use the terminal and connected the MCP. All the steps can be found here https://www.sara-taher.com/tools/google-knowledge-panel-mcp how-to-install . You can also check the github repo directly here. https://github.com/theseoriddler/google knowledge panel mcp You will need to have an API key for Google Knowledge Graph API. Auditing my entity with the new Knowledge Panel MCP I think this was the most fun part. I asked claude to “analyze my knowledge panel “sara taher” and how strong it is compared to other entities under the same name” Using the MCP, it found that there are 20 entities sharing the same name, and to my surprise the most prominent one is NOT me Claude use the tools in the MCP, did the research and came back with a summary table. WOW I always thought the problem was the noise around my entity having other “sara taher” people, but apparently the other Sara has a some strong signals that I have to compete with 😄 Answers Vs. Data If you’ve been following my work, I’m a big fan of python scripts. Could this task been done with one? absolutely and I’m sure there are cases where a python script would have been more ideal. But what using an MCP did here is help me have a conversation about the data, and draw conclusions or possible areas to explore. A python script using the API will help you get the data, you will be in full control, you will do the analysis all by your own. I love that and I think that this is very powerful. But an MCP comes in handy if you don’t want to build a new python script every time you have a different use case for this API, if you want to have a conversation or a brainstorming session about the data, or looking for a quick check rather than invest time in building and testing a python script. Overall I highly recommend you play around with MCPs, and this MCP of course 😄😄😄 And That’s a Wrap Almost 😄 I loved every bit of this and I think MCPs will have a special place in my workflows, but also python scripts are still in my heart 😄 I think in this era, we should try everything at least once ideally more to get a good taste of the experience before deciding if it works for us or not and for what use cases. If you try this MCP and have any questions, my LinkedIn DMs https://www.linkedin.com/in/sara-seo-specialist/ are wide open That’s that for today folks and see you in the next newsletter Support the Riddler - Sign up for my newsletter if you’re not already. Pssst, you can also become a paid subscriber - Share the newsletter and invite your friends to signup. Help me reach 2k signups on Substack by end of 2026 please 🙂 - Provide feedback on how I can make this newsletter better - Buy me coffee https://buymeacoffee.com/sarataher?ref=sara-taher.com . - If you’re an SEO tool or an SEO service provider, consider sponsoring my newsletter. I’m also open to other partnership ideas as well. Disclaimer: LLMs were used to assist in wording and phrasing this blog.