Model or dataset
Ryze-AI-Adgent/open-seo-mcp-skills avatar
Ryze-AI-Adgent/open-seo-mcp-skills

Eight Markdown skills, one hosted connector, and a free claim about the code

Open-source SEO + GEO skills for Claude on your real GSC/GA4/ads data. Free MCP: claude mcp add ryze --transport http https://connector.get-ryze.ai/mcp

9,958 stars399 forksShellMIT

At a glance

What is it?
open-seo-mcp-skills is an MIT collection of eight Markdown skills that give Claude SEO actions backed by Search Console, GA4 and ads data. The skills are free and the repository is small; the data path runs through a hosted connector, and the readme's cost arguments are the vendor's own numbers.
Who is it for?
Suit this repository to someone who already has Search Console and GA4 connected, wants an agent to read their own numbers instead of scraped estimates, and is willing to sign in to a hosted connector to make that possible. Three things to check first.
Can I use it commercially?
Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
Is it still maintained?
Yes. The repository last received commits 19 days ago.
What is it written in?
Mainly Shell, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.

From daily snapshots

Star history

  • Ryze-AI-Adgent/open-seo-mcp-skills9,958 ★ · +9,183
Star history
Repository2026-09-112026-11-15
Ryze-AI-Adgent/open-seo-mcp-skills77510k

9,183 Stars in the last 65 days, measured from the daily GitHub snapshot. Compare with other projects

Editorial analysis

The payload is eight Markdown files, and install.sh explains the language

The top level of the repository is six entries: a licence, the readme, an image, an install script, a skills directory, and the plugin manifest directory Claude Code reads. The recorded primary language is Shell, which comes from that single install script rather than from anything the project computes. What the project actually ships is Markdown under skills/, and the readme says so directly, because the skills can be pasted into any client rather than needing a runtime. That is the design decision worth naming. An agent skill here is instructions in a file, so the same text works in Claude Code, in Claude Desktop, and in a client that has no plugin system at all.

Three commands, one URL, and a copy fallback

The quick start is three lines, and the first is the connector rather than the skills:

bash
claude mcp add ryze --transport http https://connector.get-ryze.ai/mcp
claude plugin marketplace add Ryze-AI-Adgent/open-seo-mcp-skills
claude plugin install open-seo-mcp-skills@ryze

After that the interaction is a sentence, such as asking for an SEO audit on a given site. The page notes that the plugin bundles the connector, so in Claude Code the first line is optional. There is a second route through Claude's own interface, Settings then Connectors then Customize, where a custom connector is added by name and URL, followed by signing in, picking a workspace and connecting the Search Console, GA4 and ads accounts once. And there is a third route with no plugin system involved at all:

bash
git clone https://github.com/Ryze-AI-Adgent/open-seo-mcp-skills && cp -r open-seo-mcp-skills/skills/* ~/.claude/skills/

That third route is the one to keep in mind if you are copying the skills into a directory you already manage yourself, because it needs nothing from the vendor's plugin machinery and leaves you with plain files. The repository has no GitHub releases to install from, so a clone is also the only versioned way to get the current set.

The data source column is the actual design of the skill set

The skills table is the most informative page in the readme, because its third column is what each skill reads. Four of the eight work from accounts you already own: the site audit covering indexation, click-through rate anomalies and decaying pages reads Search Console and GA4, rank tracking compares position movers between any two periods and reads Search Console, AI visibility reads GA4 referral reports, and the ads comparison crosses Search Console with Google Ads to find queries you pay for that you already rank for. Three reach outside for what Google will not give you, taking competitor keywords, backlinks and result pages from DataForSEO, with a content brief built from the same place. Keyword research is the one that mixes both, starting from Google Ads keyword data and widening with DataForSEO.

Free is a claim about the tool, not about the data path

The licensing statement is short: no subscription for the tool, no markup on API calls, and the skills are MIT, free to take, change and ship. That is true of the files. The architecture is where the reading gets interesting, because the skills do not talk to Google or to DataForSEO directly. They talk to a Model Context Protocol server on the vendor's own domain, which is where the accounts are connected and where the data is assembled. So the price of the data path is not stated in this repository beyond saying it runs through your Ryze workspace.

The cost comparison is the vendor's own arithmetic

The page argues that most open source SEO tools are a user interface over a data API, and that you either bring your own key and pay per request or pay a hosted subscription with a markup on every call. It puts that markup at roughly twenty-eight percent, and puts the commercial incumbents at a hundred and twenty dollars or more per month, against free for this project. Those figures appear only in the readme's tables, attributed to nobody, and the project offering the comparison is the one whose connector the skills depend on, so the table is marketing that happens to be formatted as evidence. What is checkable is the qualitative difference the table draws: estimated positions from scraped results on one side, your own Search Console positions on the other, and the same split for traffic and for AI referrals.

Tool names carry the provider, and there is a docs escape hatch

The section aimed at people extending the skills is the most useful thing in the repository. Tools on the connector are namespaced by provider and tool name, so a call reads like a provider prefix, two underscores, then the operation, for example a Search Console raw search analytics call or a Google Ads keyword ideas call. When the shape of a request is unclear, the convention is to call a documentation tool and pass the provider name, which returns the official API docs for it. The honest limitation is in the same paragraph: tool availability can vary by workspace, and every skill in this repository is expected to state its fallback. That is a stronger promise than most projects make, because it means a missing tool has a documented behaviour.

The GEO answer is one skill and one analytics report

Generative engine optimisation is handled by the AI visibility skill, which reads AI referral traffic in GA4 for ChatGPT, Perplexity, Claude and Gemini and reports which of them send traffic and which pages they cite. The page adds that the connector's workspace also tracks brand mentions inside AI answers, which is the piece GA4 cannot give you. On portability, the answer is that the same connector URL can be added to Cursor, ChatGPT or Claude Desktop as a custom connector, since the skills are Markdown. On the repository itself: there are no GitHub releases, the last push to the default branch is dated 27 October 2026, and the licence is MIT with the note that the skills and the repository are free while data access runs through your own workspace and your own connected accounts.

Editorial conclusion

Suit this repository to someone who already has Search Console and GA4 connected, wants an agent to read their own numbers instead of scraped estimates, and is willing to sign in to a hosted connector to make that possible. Three things to check first. Work out what the connector changes for you, since the skills are MIT and copyable but every data call goes through a Ryze workspace, and the free claim is about the tool rather than the data path. Treat the cost comparison table as a vendor claim: the figures on it, including a percentage markup on hosted wrappers and a monthly floor for the commercial incumbents, come from the page and are not substantiated in the repository. And read the notes for skill authors before extending anything, because tool availability is stated to vary by workspace and the project's own convention is that every skill declares its fallback. If you need offline or self-hosted data access, this design does not offer it.

Frequently asked questions

What is an SEO skill?

In this repository it is a Markdown instruction set that tells an AI client which data to fetch and what to do with it. Eight ship here, and the readme's table names the data each one reads: audits, rank tracking, AI visibility and ads comparison work from your own Search Console, GA4 and ads accounts, while keyword research, competitor gaps, backlink checks and content briefs add DataForSEO.

What does the Ryze MCP connector do?

It is a Model Context Protocol server at https://connector.get-ryze.ai/mcp that exposes Google Search Console, GA4, Google Ads keyword data, DataForSEO and optionally Ahrefs and Semrush as tools for Claude, ChatGPT or Cursor. You sign in, pick a workspace, and connect those accounts once.

Is open-seo-mcp-skills free to use?

The skills are MIT licensed and the page states there is no subscription for the tool and no markup on API calls. Data access runs through your own Ryze workspace and your own connected accounts, and the readme's own comparison puts commercial SEO subscriptions at a hundred and twenty dollars or more per month.

How do I install these SEO skills in Claude?

Add the connector with claude mcp add ryze --transport http https://connector.get-ryze.ai/mcp, add the marketplace with claude plugin marketplace add, then install the plugin. You can also clone the repository and copy the skills directory into ~/.claude/skills/, and the skills are plain Markdown that can be pasted into other clients.

What does the ai-visibility skill report?

It reads AI referral traffic in GA4 for ChatGPT, Perplexity, Claude and Gemini, and reports which of those engines send you traffic and which pages they cite. The page also says the connector's workspace tracks brand mentions inside AI answers, which GA4 does not report.

Official sources

  1. Issues
  2. License: MIT
  3. Project website
  4. README
  5. Ryze-AI-Adgent/open-seo-mcp-skills on GitHub
Add this badge to your README

If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.

Add this badge to your README

markdown
[![Hysen Labs](https://hysenlabs.com/badge/ryze-ai-adgent-open-seo-mcp-skills.svg)](https://hysenlabs.com/projects/ryze-ai-adgent-open-seo-mcp-skills)

Star history chart for your README

markdown
[![Star history](https://hysenlabs.com/chart/ryze-ai-adgent-open-seo-mcp-skills.svg)](https://hysenlabs.com/projects/ryze-ai-adgent-open-seo-mcp-skills)

Was this useful?