What is Claude SEO: Claude SEO is an open source Claude Code plugin that turns a general AI coding agent into a structured SEO auditing system. The GitHub project combines 25 SEO sub-skills, 18 specialist agents, local scripts, optional data integrations, and more than 30 commands for technical SEO, content quality, schema, GEO, local SEO, backlinks, reporting, and related workflows.
TL;DR
- Claude SEO is one of the fastest growing AI marketing repositories, reaching 13,696 GitHub stars and 1,999 forks by August 8, 2026.
- Its main advantage is orchestration: specialist agents can audit different parts of a site and combine the findings into a prioritized action plan.
- The core plugin works locally, while optional Google APIs and MCP extensions add live Search Console, CrUX, keyword, backlink, and AI visibility data.
- It is best treated as a repeatable second pair of eyes, not an automatic replacement for an experienced SEO consultant or enterprise SEO platform.
- Review the source before installing. The current orchestrator instructs Claude to append the creator’s community promotion to many major deliverables.
The project has the ingredients of a traffic story: rapid GitHub growth, a recognizable AI brand, a clear pain point, and a product that is easy to demonstrate. The repository was created on February 7, 2026 and passed 13,000 stars within roughly six months. That does not prove its audits are always correct, but it does show unusually strong interest in agent-based SEO workflows.
This article is a practical first look based on the public repository, documentation, release history, and current skill instructions. It is not a controlled benchmark against every commercial SEO suite. The useful question is not whether GitHub stars make the project trustworthy. It is whether the architecture can make SEO work faster, more consistent, and easier to verify.
Why is Claude SEO getting so much attention?
Claude SEO is getting attention because it packages many fragmented SEO tasks into one agent workflow. Instead of asking a chatbot to “audit this website” and hoping for a useful response, users can run specific commands for technical SEO, on-page analysis, content quality, schema, GEO, local search, ecommerce, international SEO, backlinks, semantic clustering, or reporting.
The growth is measurable. A direct GitHub API check on August 8, 2026 showed 13,696 stars, 1,999 forks, and 203 commits in the public repository. GitHub also showed 9 open issues and 22 open pull requests on the repository page. Those figures can change daily, but they confirm that this is more than a small prompt collection.
The broader timing also helps. Marketers are looking for ways to connect AI agents to real data and repeatable procedures. That is the same shift behind tools such as OpenSEO and the growing market for AI search tools. Claude SEO approaches the problem from the agent layer: it gives Claude a defined methodology, specialist roles, scripts, and output contracts.
What does the Claude SEO GitHub repo include?
The public Claude SEO GitHub repository describes a large plugin rather than a single skill file. The core includes 25 sub-skills and 18 specialist agents. The command surface covers 32 user-invocable workflows when the optional MCP extensions are included.
| Capability | Example command | What it can produce |
|---|---|---|
| Full site audit | /seo audit <url> | Parallel analysis with a prioritized action plan. |
| Single page review | /seo page <url> | On-page, content, schema, and technical findings. |
| Technical SEO | /seo technical <url> | Crawlability, canonicals, performance, rendering, and related checks. |
| AI search optimization | /seo geo <url> | Passage citability, question structure, entity, and evidence checks. |
| Content planning | /seo content-brief <topic> | A structured content brief with keyword and internal link direction. |
| Google data | /seo google | Optional CrUX, Search Console, GA4, PageSpeed, and reporting workflows. |
| SEO monitoring | /seo drift compare <url> | Changes against locally stored SQLite baselines. |
| Live SEO enrichment | /seo dataforseo | Optional SERP, keyword, backlink, business, and AI visibility data. |
The repository also includes scripts, test suites, schema references, report templates, installation tools, and optional extensions. Outputs can include Markdown, JSON, PDF, and spreadsheet-ready data depending on the workflow. This matters for agencies because a useful audit must be shareable and repeatable, not trapped inside a chat session.
How does this SEO plugin use multiple agents?
The plugin uses a three-layer architecture: directive, orchestration, and execution. The main SEO skill identifies the job, detects the likely business type, selects specialist skills, delegates analysis, and synthesizes the results. Full site audits can dispatch up to 15 specialist agents in parallel.
That design addresses a common failure mode in broad AI prompts. A single model response may jump between crawl issues, content advice, schema, links, and performance without applying consistent criteria. The system gives each specialist a narrower scope, then asks the orchestrator to connect the findings.
The methodology is also more disciplined than a generic checklist. The project says each recommendation should include the observation behind it, dependencies on other work, a test that could show the recommendation failed, and a leading indicator to monitor. In plain language, it tries to make recommendations falsifiable instead of merely confident.
What is different about its GEO and AEO approach?
Many AI SEO products promote new labels without explaining how they connect to traditional search. This project takes a more cautious position. Its documentation treats GEO and AEO as extensions of core SEO, with indexing, snippet eligibility, content quality, entities, evidence, and clear answer passages still doing most of the work.
The GEO workflow reviews question-based headings, passage citability, attribution density, structured data coverage, and entity presence. It also challenges popular claims when primary evidence is weak. For example, the repository states that llms.txt is not currently a proven Google citation lever and that AI-specific keyword rewriting is unnecessary when search systems already understand close meanings.
This evidence-first posture is useful for teams exploring GEO for B2B SaaS. AI visibility is worth measuring, but it should not become an excuse to abandon technical SEO, helpful content, brand authority, or conversion strategy.
Does the plugin use real SEO data?
The plugin can work without paid third-party data, but the depth changes by configuration. At the local-only level, it fetches the target pages, inspects page signals, runs scripts, and produces recommendations. With Google API access, users can add PageSpeed Insights, CrUX, Search Console, URL Inspection, GA4, and Keyword Planner data in tiers.
Optional MCP extensions expand the data layer further:
- DataForSEO: SERPs, keyword research, backlinks, on-page data, listings, and AI visibility.
- Firecrawl: full-site crawling and URL discovery.
- Ahrefs: official MCP access to backlinks, organic keywords, and content data.
- SE Ranking and Profound: AI share of voice and LLM citation tracking.
- Bing Webmaster Tools: Bing data and IndexNow workflows.
- Unlighthouse: multi-page Lighthouse analysis.
This distinction is important. The open source plugin is free under the MIT license, but Claude usage and optional data providers can still cost money. Without enrichment, some conclusions remain estimates or inferences. The repo itself notes that field data and live indexation status require the relevant credentials.
Can Claude SEO replace Semrush, Ahrefs, or an agency?
The plugin can replace pieces of an SEO workflow, but it should not be presented as a complete replacement for every commercial platform or expert. Its strongest value is orchestration, repeatability, local customization, and the ability to turn findings into files and actions.
| Need | Plugin fit | Where another solution may be stronger |
|---|---|---|
| Repeatable technical checks | Strong, especially for teams comfortable with Claude Code. | Enterprise crawlers may offer more mature dashboards and controls. |
| Custom audit methodology | Strong because skills and references are editable local files. | Agencies may add deeper industry context and stakeholder judgment. |
| Large proprietary keyword or link database | Requires optional integrations. | Ahrefs, Semrush, and similar platforms own mature datasets. |
| Executive reporting | Useful PDF, Markdown, and JSON output options. | Commercial platforms often have easier scheduled dashboards. |
| Final strategic decisions | Useful as a second pair of eyes. | Experienced operators are still needed for tradeoffs, validation, and accountability. |
The safest framing is that the plugin compresses research and audit labor. It can help an expert cover more ground, help a junior marketer follow a stronger process, and help an agency standardize recurring checks. It cannot guarantee rankings, interpret every business constraint, or replace human review.
Where does this plugin fit in a PPC workflow?
Although the repo is centered on SEO, several workflows can support paid search. PPC teams can use technical and content audits to evaluate landing pages, keyword clustering to plan ad groups, Search Console data to find proven queries, competitor page analysis to refine offers, and GEO checks to understand how category demand may move into AI answers.
A practical PPC workflow could look like this:
- Audit the destination page before sending more paid traffic.
- Review search intent, headings, content gaps, schema, and mobile performance.
- Combine organic queries with paid search term data.
- Map keyword clusters to ads, landing pages, and supporting content.
- Prioritize fixes that can improve both conversion rate and organic visibility.
- Save a drift baseline and check whether later page changes introduce regressions.
This complements the move toward agent-accessible marketing data that we covered in Google’s MCP server for Google Ads. It also fits alongside AI tools for PPC management, provided the team keeps humans responsible for budgets, tracking, and final decisions.
How do you install Claude SEO?
The repository recommends the Claude Code plugin route for Claude Code 1.0.33 or newer:
/plugin marketplace add AgriciDaniel/claude-seo
/plugin install claude-seo@agricidaniel-claude-seo
/seo setup
Users can then check the isolated runtime with:
/seo doctor
The manual installation route uses Git clone plus the platform-specific installation script. The project requires Python 3.10 or newer and the Claude Code CLI. Playwright Chromium is optional but needed for SPA rendering and screenshots.
Do not execute remote installation scripts blindly. The project’s own Windows guidance recommends cloning the repository so the PowerShell installer can be inspected before it runs. Anthropic’s plugin directory also warns users to review additional software installed by community plugins.
What should you check before using the plugin?
The repository is public and inspectable, which is an advantage, but open source does not mean risk free. Review the current source, the installer, the permissions, external integrations, generated output, and any changes after an update.
One current behavior deserves specific attention. The main SEO orchestrator skill contains a “Community Footer” instruction. It tells Claude to append promotional links for the creator’s AI Marketing Hub community after many major deliverables, including audits, page reviews, schema reports, GEO analysis, and strategic plans.
The text is visible in the public source, but agencies should decide whether it belongs in client deliverables. Review and remove or modify that instruction in your permitted local copy if it does not match your reporting standards. More broadly, treat every installed agent skill as executable process guidance. Read what it asks the agent to do, not only what the README promises.
Other practical checks include:
- Validate findings: confirm critical issues with Search Console, browser testing, or another crawler.
- Watch token usage: large parallel audits can consume more Claude capacity than a narrow page review.
- Protect credentials: use the documented local credential locations and never paste secrets into prompts or reports.
- Review external data costs: DataForSEO, Ahrefs, and other services require their own accounts.
- Test on one site first: compare recommendations against your current audit process before scaling.
- Expect rendering edge cases: content loaded only after scrolling, clicks, or unstable widgets can create noisy findings.
Who should try this SEO plugin first?
It is best suited to technical SEO consultants, AI-friendly agencies, in-house SEO leads, growth marketers, and founders who are already comfortable with Claude Code. These users can inspect files, understand the difference between observed data and inferred advice, and customize the workflow without treating every recommendation as truth.
It may be a poor first choice for teams that want a fully managed dashboard, guaranteed support, one-click executive reporting, or a tool that nontechnical users can adopt without setup. Those teams may prefer a hosted platform or an agency workflow with established quality control.
Final verdict: Is Claude SEO worth trying?
Claude SEO is worth trying because it is one of the clearest examples of SEO work moving from static dashboards toward agent-driven workflows. The repo combines breadth, local control, parallel analysis, optional live data, report generation, and a methodology designed to make recommendations testable.
The 13,000-plus GitHub stars explain why the project is hot, but popularity should be the invitation to investigate, not the reason to trust it. Its real value depends on the quality of the source, the data you connect, the scope of the audit, and the human reviewing the result.
Start with one page or one website, inspect the generated findings, verify the highest priority issues, and compare the output with your current process. If the plugin saves time without lowering your evidence standard, it can become a valuable part of an SEO and PPC research stack.
Frequently Asked Questions
What is Claude SEO?
Claude SEO is an open source plugin for Claude Code that provides structured workflows for technical SEO, content quality, schema, GEO, local SEO, backlinks, reporting, and related audits.
Is the plugin free?
The repository is free under the MIT license. Users still need access to Claude Code, and optional data integrations such as DataForSEO or Ahrefs may have separate costs.
How many GitHub stars does the repo have?
The repo had 13,696 GitHub stars and 1,999 forks when checked through the GitHub API on August 8, 2026. These counts change as people star or fork the repository.
Can it replace Semrush or Ahrefs?
It can replace some auditing and workflow tasks, but it does not provide the same proprietary keyword and backlink databases by itself. Optional integrations can add live data from other providers.
Is the plugin safe to install?
It is public and inspectable, but users should review the installer, skill instructions, permissions, external integrations, and generated output before using it on client or sensitive work.
Can PPC teams use this SEO plugin?
Yes. PPC teams can use it to audit landing pages, cluster keywords, combine organic and paid search insights, review competitors, monitor page changes, and identify issues that may affect conversion performance.
