My GEO MCP verdict is that the four real GEO MCP servers on GitHub are useful audit toolboxes, but none of them is a complete GEO workflow, so my own Agent OS ranks first on this list.
A GEO MCP is a Model Context Protocol server that gives an AI agent generative engine optimisation tools, such as AI crawler checks, schema checks and citation-readiness scores.
Yes, Agent OS is my own product, so this is my own list, and I'll show you exactly how I scored everything.
I rated every option feature by feature from its own README and the GitHub data I pulled on 19 October 2026.
I haven't run the four community repos myself, so these are reference ratings rather than hands-on test scores, and I'll say so wherever it matters.
If you'd like a strategist to look at your AI visibility instead, book a free one-to-one GEO, AEO AI audit and get a custom game plan.
GEO MCP Ranking And Scores At A Glance
Here's the ranking table with my scores, so you get the verdict in five seconds.
| Rank | Option | Score | Best for |
|---|---|---|---|
| 1 | Agent OS (my own, inside the AI Profit Boardroom) | 9/10 | People who want the full GEO workflow, from Search Console data to published case-study pages. |
| 2 | Hermes Agent on its own | 8/10 | People who want the same workflow for free without my dashboard. |
| 3 | tmolavi/mcp-geo-server | 7/10 | Deep three-pillar page audits with an evidence label on every finding. |
| 4 | jpurnell/GeoSEOMCP | 6.5/10 | Mac users who want 29 repeatable offline scoring tools. |
| 5 | SEOSiri-Official/aeo-geo-mcp | 5/10 | Quick llm.txt and schema density checks. |
| 6 | henu-wang/geoscore-mcp | 4/10 | Hosted scans, if the GEOScore service is reachable. |
The top two are scored as workflows I actually show in my video, while the four repos are scored from their READMEs.
I'm putting that difference in plain sight so you can weigh the numbers properly.
How I Scored Every GEO MCP
I used five criteria, and I weighted them equally.
| Criterion | What I looked for |
|---|---|
| Tool coverage | How many of the main GEO checks it covers, such as crawler access, schema, citability and content quality. |
| Documentation and honesty | Whether the README explains every tool, how to install it and what its scores do and don't mean. |
| Freshness | When the code was last pushed and whether the install path worked when I checked. |
| Ease of setup | How hard it is to get running for a normal operator. |
| Distance to results | How close it gets you to a published page that can actually be named in an AI answer. |
The last criterion is why my top two beat every repo.
A GEO audit tool can score a page, but it can't pick your next topic or write your case study.
What A GEO MCP Is, In One Reference Section
MCP stands for Model Context Protocol, the open standard for connecting AI apps to outside tools.
The MCP introduction explains it, and both Claude Code and Hermes Agent support it.
GEO stands for generative engine optimisation, which is the work of getting your business named in AI answers from Google AI, Gemini or ChatGPT.
Put the two together, and a GEO MCP is a server of GEO tools that your agent can call by name.
🔥 Want my exact Best GEO MCP Servers Rated setup, step by step? Inside the AI Profit Boardroom you get step-by-step video trainings, my installable Agent OS and four coaching calls a week with 3,400+ members building real automations. → Join the AI Profit Boardroom
#1 Agent OS Reviewed: 9/10
Agent OS is my own dashboard that runs a crew of AI agents, such as Hermes and Claude, under one shared memory.
It isn't on GitHub, because it's only available inside the AI Profit Boardroom as an installable zip with guides.
It scores highest because it runs the full workflow I show in my video, not just the audit step.
| Feature | What it does in the workflow |
|---|---|
| Search Console analysis | I export my Google Search Console data to CSV and ask the agent for keyword and topic ideas to rank inside AI search engines. |
| Report artifact | The agent builds a report showing clicks, brand share of clicks, non-brand click-through rate, the month's clicks and a verdict on the best pages. |
| SEO skill | A checklist of roughly 100 points that writes everything from a case-study angle. |
| Publishing | The agent can publish to my sites once I give it access, such as WordPress or Netlify. |
| Search OS | Inside the Boardroom there's a search OS that finds GEO keywords from your own Search Console data. |
I lose a point because it isn't free, and because you still have to supply the real case studies and fact-check every page.
#2 Hermes Agent On Its Own Reviewed: 8/10
In my video I say you can run the same process with Hermes Agent directly, without my Agent OS.
Hermes Agent is free and open source, and I use OpenCode as the free brain.
When I recorded, DeepSeek V4.1 Flash was free inside OpenCode.
It scores slightly lower because you build the dashboard, memory and SEO skill yourself.
Hermes also supports MCP, so you can add any GEO MCP below to it later if one earns its place.
#3 tmolavi/mcp-geo-server Reviewed: 7/10
The mcp-geo-server repo is a Python MCP server by Taqi Molavi, with an MIT licence and 3 stars on 19 October 2026.
Its README describes it as a thin adapter that exposes the SAGE audit engine to agents like Claude Desktop, Cursor and Antigravity.
| Feature | What the README says it does |
|---|---|
| Full audits | It runs a three-pillar audit of a live URL or of raw HTML you paste in. |
| Technical SEO pillar | It checks DOM hygiene, headings, canonicals, header policies and AI crawler rules for GPTBot, ClaudeBot and PerplexityBot. |
| Entity AEO pillar | It checks JSON-LD depth, sameAs links and how many direct answers a page contains. |
| GEO pillar | It simulates retrieval chunking, measures semantic density and gives a Citation Survival Proxy score. |
| Citation readiness | It checks fact density, source attribution and numbers that can be verified. |
| llms.txt | It generates an llms.txt file for a domain. |
| Evidence levels | It labels every finding from E0 for facts to E5 for hypotheses. |
I like the honesty here, because the README says its proxy scores are diagnostics and not citation probabilities.
It loses points because the README installs it from PyPI, but I couldn't find the package on the public PyPI index on 19 October 2026, so you may need the repo source.
#4 jpurnell/GeoSEOMCP Reviewed: 6.5/10
The GeoSEOMCP repo is a Swift MCP server with 0 stars on 19 October 2026, and its README says it's MIT licensed.
The README lists 29 tools, which is the widest coverage of any repo here.
| Tool group | Tools in the group |
|---|---|
| Citability | Two tools that score a passage or a whole page for how quotable it is. |
| Content analysis | Four tools for readability, content statistics, an E-E-A-T score and page-type benchmarks. |
| Crawler access | Three tools that parse robots.txt and score AI crawler access. |
| Schema | Four tools that validate JSON-LD, audit sameAs coverage, score completeness and generate templates. |
| Technical SEO | Five tools for a technical score, security headers, headings, meta tags and server-side rendering. |
| Brand and platform | Four tools for brand authority and platform presence. |
| llms.txt | Two tools that validate llms.txt and sort URLs for it. |
| Composite and utility | Five tools for a weighted GEO score, finding priorities, business type, syllables and pronoun density. |
The README says no tool fetches a URL, so you or your agent supply the text and the scores stay offline and repeatable.
It was pushed on 17 October 2026, so it's actively maintained.
It loses points on setup, because it needs Swift 6.2 or newer and macOS 14 or newer.
#5 SEOSiri-Official/aeo-geo-mcp Reviewed: 5/10
The aeo-geo-mcp repo is a local-first Python MCP server by Momenul Ahmad of SEOSiri, with an MIT licence and 0 stars on 19 October 2026.
Its README lists 10 tools, including an llm.txt compliance audit, a security.txt check, a GEO AI-readiness score from 0 to 100, answer-card extraction, a domain authority rank lookup and a schema density check.
It also has a content stickiness tool that compares GA4 session duration with bounce rate.
It scores lower because the README is short and gives one line of explanation per tool.
📞 Want a strategist's verdict on your own site? Book a free AEO SEO strategy session and my team will look at your site, your competitors and the keywords you're closest to winning. → Get your free one-to-one GEO, AEO AI audit
#6 henu-wang/geoscore-mcp Reviewed: 4/10
The geoscore-mcp repo is a JavaScript MCP server for the GEOScore platform, with 1 star on 19 October 2026, and its README says it's MIT licensed.
Its README lists eight tools, plus guide resources and three ready-made prompts.
| Tool | Needs an API key? |
|---|---|
| A domain scan with a score, a grade and 11 checks. | No, according to the README. |
| A deep scan with an implementation report. | Yes, a Pro key. |
| A fix generator for llms.txt, schema, meta, robots and sitemap. | Yes. |
| A domain profile and score history. | No. |
| A side-by-side comparison of two domains. | No. |
| An AI search citation check. | Yes, a Pro key. |
| A saved report lookup. | No. |
| A tool to create a free API key. | No. |
It scores lowest because the code hasn't been pushed since 5 March 2026.
When I tried geoscoreai.com on 19 October 2026 it returned an error, and I couldn't find the package on the public npm registry that day.
🚀 Ready to put Best GEO MCP Servers Rated to work in your business? Inside the AI Profit Boardroom you get step-by-step video trainings, my installable Agent OS and four coaching calls a week with 3,400+ members building real automations. → Join the AI Profit Boardroom
Feature Matrix: Which GEO MCP Covers What
This table compares the four repos on the checks people ask about most, using only what each README says.
| Check | mcp-geo-server | GeoSEOMCP | aeo-geo-mcp | geoscore-mcp |
|---|---|---|---|---|
| AI crawler access in robots.txt | Yes, in the technical pillar. | Yes, three crawler tools. | Not listed. | Robots fixes are listed. |
| Schema validation | Yes, in the entity pillar. | Yes, four schema tools. | Yes, a density check. | Schema fixes are listed. |
| Passage citability | Yes, citation readiness. | Yes, two citability tools. | Answer-card extraction. | Not listed. |
| llms.txt | Generates it. | Validates it. | Audits it. | Generates it with a key. |
| Fetches live URLs | Yes, or takes raw HTML. | No, you supply the text. | Not stated clearly. | Yes, through the hosted API. |
| Needs an outside service | No. | No. | No, apart from the rank lookup. | Yes. |
The llms.txt Question, Settled
All four of these repos check or generate an llms.txt-style file, so it's worth being clear about it.
Google's AI features and your website page says you don't need new machine-readable files, AI text files or special schema to appear in AI Overviews or AI Mode.
It says normal SEO best practice still applies and there are no additional requirements.
So I'd never rank a GEO MCP higher just because it scores llms.txt.
How To Install Any GEO MCP, The Official Way
In Claude Code, Anthropic's MCP docs use claude mcp add --transport http <name> <url> for a remote server.
For a local server, the docs use claude mcp add <name> -- <command> [args...], with the server's own launch command after the double dash.
You check the result with claude mcp list, claude mcp get <name> or the /mcp command inside Claude Code.
In Hermes Agent, the Hermes MCP docs say you add the server under mcp_servers in ~/.hermes/config.yaml.
A local server takes a command and args, and a remote server takes a url.
The Hermes docs also let you limit a server's tools with tools: include:, which I'd do for any GEO MCP.
Safety Verdict: Review Every Server First
I'd mark every option down to zero if you install it without reviewing it.
Anthropic's docs say to verify you trust each server before connecting it, and they warn that servers that fetch outside content can expose you to prompt injection.
The Hermes docs say you should read what a catalogue entry installs and runs before you install it.
So read the source, check what each tool can reach, register only the read-only tools you need and test on a throwaway project first.
📞 Rather not test tools on your live site? Claim a free strategy session and we'll show you why competitors appear in AI answers more often than you do. → Book the free call
Does The Winning Workflow Pass Google's Guidance?
Google updated its page on using generative AI content on 1 October 2026.
It says it's "critical to manually factcheck and review all AI-generated content" before publishing, including titles, descriptions, structured data and alt text.
My workflow passes that test only if a human checks every page, and that's why my SEO skill writes from real case studies.
A page built on your own test is easier to check, because you were there when it happened.
Here's a second video where I show these ideas working inside Google's AI Mode.
The Evidence Behind My Top Two
These are my own results from my own screens in the video, not a promise.
- For "best geo expert", my content shows in the Google AI answer at the top and ranks first organically.
- In AI Mode, asking for the best GEO experts to follow in 2026 lists me first and links to my sites.
- Perplexity also lists me among GEO experts, even though I don't really use Perplexity.
- One of my sites went from zero to more than 1,300 clicks a day, and two other sites are on the same path.
None of that came from a GEO MCP, which is the clearest reason the repos sit below the workflow.
Final Verdict On GEO MCP Servers
If you want results, start with the Search Console and SEO skill workflow, either inside my Agent OS or with Hermes Agent for free.
If you want an audit toolbox on top, tmolavi/mcp-geo-server has the most complete README and GeoSEOMCP has the widest offline tool set.
SEOSiri's server is fine for quick checks, and I'd wait on GEOScore until its service and code look active again.
Every one of them is a small community project, so review the code before you trust it.
🔥 Want the Agent OS and SEO skill that ranked first here? Inside the AI Profit Boardroom, you get the Agent OS, my SEO skill checklist, the AI SEO automation section, four coaching calls a week and 3,400+ members. → Get access here
Related Reading
- See four real GEO examples rated for repeatability.
- Find out how to run a free GEO audit.
- Read my AEO audit walkthrough.
- Compare generative engine optimisation with SEO.
Also On Our Network
- 🌐 Set up a GEO MCP in Claude Code or Hermes step by step
- 🌐 Five practical ways agencies use a GEO MCP
- 🌐 A quick-start GEO MCP guide for an Agent OS
- 🌐 How the models behind GEO MCP agents score
FAQ: GEO MCP Ratings
What is the best GEO MCP server?
On my own list, Agent OS ranks first because it runs the full Search Console and SEO skill workflow.
Among the community repos, tmolavi/mcp-geo-server scores highest on my README-based ratings.
How did you score the GEO MCP servers?
I scored each one out of 10 on tool coverage, documentation, freshness, ease of setup and distance to results.
I used each repo's own README and GitHub data from 19 October 2026, so these are reference ratings rather than hands-on test scores.
Are GEO MCP servers popular?
No, because on 19 October 2026 the four repos I found had between zero and three GitHub stars.
Do GEO MCP servers need llms.txt?
Several of them check or generate llms.txt, but Google says you don't need AI text files or special schema to appear in AI Overviews or AI Mode.
Should I review a GEO MCP before installing it?
Yes, because an MCP server runs code and acts for your agent.
Read its source, check what each tool can reach and only connect servers you trust.
Can I do GEO without a GEO MCP?
Yes, and the workflow in my video uses an agent, a Search Console CSV export and an SEO skill with no GEO MCP at all.
📺 Video notes + links to the tools 👉
🎥 Learn how I make these videos 👉
🆓 Get a FREE AI Course + Community + 1,000 AI Agents 👉
About Julian
I'm Julian Goldie, an SEO entrepreneur, author and founder of the AI Profit Boardroom, which has 3,400+ members.
I help business owners scale with AI agents, automation and SEO.
- I started Goldie Agency on my own, and it is now a 7-figure SEO and link-building agency with 70+ people.
- I've grown my YouTube channel to 400,000+ subscribers.
- I wrote "Link Building Mastery".
- My Udemy courses have taught 29,000+ students.
→ Get my best AI training inside the AI Profit Boardroom
Pick the workflow first, book a free strategy session if you want a strategist's view, and only then choose the GEO MCP that fills a real gap.
