HOME SKILLS BLOG GITHUB
// TUTORIAL

How to Run a Free SEO Audit With Claude SEO (Step-by-Step Guide)

By Daniel Agrici | | Last updated: | 8 min read | TUTORIAL SEO AUDIT FREE
How to run a free SEO audit with Claude SEO step-by-step guide

An SEO audit is a structured review of the technical, content, and structured-data conditions that affect crawling, indexing, usability, and search eligibility. Claude SEO provides an open-source terminal workflow for that review. Completion time depends on site size, network access, rendering, enabled integrations, and the analyses selected.

LISTEN // 6:44 NARRATION

What is Claude SEO and why should you use it?

Claude SEO is an MIT-licensed command-line tool that runs inside Claude Code. Version 2.2.4 contains 25 subskills, 18 specialist agents, 32 user-facing commands, and a release suite of 410 passing tests. Optional APIs have their own accounts and costs.

Claude SEO
$0
Open source, terminal-based, MIT license

Unlike a managed browser dashboard, Claude SEO runs through your terminal and routes work to relevant specialist agents. Its weighted 0-100 health score is an internal prioritization heuristic. It is not a Google score, a ranking signal, or a guarantee that a recommendation will improve traffic.

SEO HEALTH SCORE - CATEGORY WEIGHTS Content Quality 23% Technical SEO 22% On-Page SEO 20% Schema/Structured Data 10% Core Web Vitals 10% AI Search (GEO) 10% Image Optimization 5% Weighted scoring across 7 categories totaling 100 points

Step 1: Install Claude SEO

Before you begin, make sure you have two prerequisites installed:

  • Python 3.10+ - check with python3 --version
  • Git - required for manual installation and repository workflows
  • Claude Code CLI - Anthropic's terminal-based AI assistant

The recommended plugin flow installs Claude SEO first, then provisions its isolated runtime explicitly:

/plugin marketplace add AgriciDaniel/claude-seo
/plugin install claude-seo@agricidaniel-claude-seo
/seo setup
/seo doctor

/seo setup creates the managed environment. /seo doctor performs a read-only check. Manual installation remains available for users who prefer to inspect the repository path first.

macOS or Linux

git clone --depth 1 https://github.com/AgriciDaniel/claude-seo.git
bash claude-seo/install.sh

Windows

git clone https://github.com/AgriciDaniel/claude-seo.git
powershell -ExecutionPolicy Bypass -File claude-seo\install.ps1

The installer clones the repository and registers the skill with Claude Code. You should see a confirmation message when it completes.

Step 2: Run your first full audit

The full site audit routes technical, content, structured-data, performance, image, sitemap, and AI-search checks to relevant specialist agents.

/seo audit https://yoursite.com
Running a Claude SEO command in the terminal

This command summarizes findings into seven weighted score groups and a prioritized report.

CategoryWeightWhat it checks
Content Quality23%E-E-A-T signals, depth, readability, originality
Technical SEO22%Crawlability, indexability, security headers, JS rendering
On-Page SEO20%Title tags, meta descriptions, heading hierarchy, keyword placement
Schema/Structured Data10%JSON-LD detection, validation against Google requirements
Core Web Vitals10%LCP, CLS, INP (replaced FID in 2024)
AI Search Readiness10%GEO score, AI crawler access, passage-level citability
Image Optimization5%Alt text, file sizes, lazy loading, responsive images

Run time varies with site size, network conditions, browser rendering, and optional integrations. Record the audited URL set and enabled data sources when comparing results over time.

Step 3: What does your health score mean?

Claude SEO assigns every finding a priority level so you can review higher-risk issues first. Some checks reference Google Search Essentials or the published Core Web Vitals thresholds. The combined score and category weights remain Claude SEO's vendor-specific heuristic.

SCORE INTERPRETATION GUIDE <60 Poor 60-69 Below Avg 70-79 Needs Work 80-89 Good 90-100 Exceptional Internal prioritization bands, not Google ranking thresholds Validate every recommendation against evidence and observed outcomes

Each finding comes with a specific recommendation. For example, a site scoring 64/100 might show: "Multiple H1 tags detected on homepage (Critical), thin content on /about page (High), missing security headers (Medium), truncated title tags on 3 pages (Medium)."

Step 4: Run targeted deep dives

After the full audit, drill into specific areas that need evidence or deeper analysis. Version 2.2.4 exposes 32 user-facing commands across the toolkit.

Technical SEO deep dive

/seo technical https://yoursite.com

This checks 9 sub-categories: crawlability, indexability, security, URL structure, mobile optimization, Core Web Vitals (including INP), structured data, JavaScript rendering, and IndexNow protocol support.

Content quality assessment

/seo content https://yoursite.com

Evaluates your content against the September 2025 Quality Rater Guidelines for E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). It checks author attribution, source citations, content depth, and whether your content demonstrates first-hand experience.

Schema markup validation

/seo schema https://yoursite.com

Detects existing Schema.org markup, validates it against Google's requirements, and generates corrected JSON-LD if issues are found. This is particularly valuable because schema errors are invisible to visitors but can prevent rich results in search.

AI search readiness (GEO)

/seo geo https://yoursite.com

The GEO review checks search eligibility, relevant crawler controls, content clarity, evidence, authorship, and platform measurement. Its score is a heuristic. Google states that AI Overviews and AI Mode use established SEO foundations and do not require special markup, an llms.txt file, or AI-specific rewrites.

Step 5: How do you prioritize fixes?

Work through findings in order: Critical first, then High, Medium, and Low.

CRITICAL Same day HIGH This week MEDIUM This month LOW Ongoing Re-run /seo audit after each batch of fixes to measure improvement (10-20 point gains typical)
  1. Critical fixes (same day): Multiple H1 tags, broken canonical URLs, blocked crawlers, missing SSL
  2. High priority (this week): Thin content pages, missing alt text, no schema markup, blocked AI crawlers
  3. Medium priority (this month): Title tag optimization, internal linking gaps, image compression
  4. Low priority (ongoing): Minor meta description improvements, additional schema types, performance fine-tuning

After making fixes, re-run the audit to measure improvement:

/seo audit https://yoursite.com

Compare your new score against the baseline. Most sites see a 10-20 point improvement after addressing critical and high-priority findings.

Step 6: Extend with optional integrations

Claude SEO's core functionality needs no API keys. But two optional extensions unlock additional capabilities.

DataForSEO integration

/seo dataforseo

This connects to the DataForSEO API for real-time search engine results, keyword volume and difficulty metrics, and competitor backlink profiles. Requires a DataForSEO API key (pay-as-you-go pricing).

AI image generation

/seo image-gen "professional hero image for SEO audit blog post"

Uses Google Gemini via the nanobanana-mcp server. Requires a GOOGLE_AI_API_KEY (free at aistudio.google.com).

MCP server integrations

Claude SEO also works with official MCP servers from Ahrefs, Semrush, Google Search Console, and PageSpeed Insights for even deeper analysis.

ALL 16 CLAUDE SEO COMMANDS CORE ANALYSIS /seo auditFull site audit (up to 15 specialists) /seo pageSingle-page deep analysis /seo technical9 technical sub-categories /seo contentE-E-A-T quality assessment STRUCTURED DATA & MEDIA /seo schemaSchema.org detection + JSON-LD /seo imagesAlt text, sizes, lazy loading /seo sitemapXML sitemap analysis AI & STRATEGY /seo geoAI search readiness (GEO) /seo planIndustry-specific strategy /seo localLocal SEO + GBP analysis /seo hreflangInternational SEO audit /seo programmaticPages at scale /seo competitor-pagesX vs Y comparison pages
DA
Daniel Agrici
AI Automation Specialist. Creator of Claude SEO, Claude Blog, and Claude Banana. GovTech Hackathon Moldova 1st place winner. GitHub / YouTube / LinkedIn
// FAQ

FREQUENTLY ASKED QUESTIONS

Claude SEO is MIT-licensed and has no paid product tier. You still need Claude Code, and optional services such as DataForSEO or Gemini use their own accounts, quotas, and pricing. Those external costs are not Claude SEO fees.
Claude SEO provides a terminal-based audit and recommendation workflow. It is not a replacement for the proprietary backlink indexes, keyword databases, historical tracking, or managed interfaces in paid platforms. Choose tools according to the data and workflow you need.
It can review public pages that its fetchers or browser can access while respecting site controls. Private, authenticated, blocked, or technically unavailable content requires authorization and additional access.
There is no universal audit schedule. Run an audit after significant releases, migrations, template changes, or search-policy updates, and choose a recurring cadence that matches how often your site changes.
Yes, when the platform exposes accessible public HTML or can be rendered by the configured browser. The audit evaluates the output it can retrieve, so blocked, authenticated, or incomplete rendering may limit the result.
// RELATED

KEEP READING

CLAUDE SEO VS AHREFS VS SEMRUSH → FREE SEO AUDIT TOOLS ROUNDUP →

RUN YOUR FREE AUDIT
IN 30 SECONDS.

$
git clone --depth 1 https://github.com/AgriciDaniel/claude-seo.git && bash claude-seo/install.sh
VIEW ON GITHUB ALL 25 SKILLS >