The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the Seo listing page.
An MCP (Model Context Protocol) server that gives AI agents a suite of SEO analysis tools. Analyze any public webpage for meta tags, heading structure, broken links, keyword density, page speed, and sitemap health — all without external API keys.
| Tool | Description |
|---|---|
seo_meta_analyze | Audit meta tags, Open Graph, and Twitter Card tags with scoring |
seo_heading_check | Validate H1-H6 hierarchy, detect skipped levels and duplicates |
seo_link_check | Find broken links, redirects, and missing anchor text |
seo_keyword_density | Measure keyword frequency and check target keyword placement |
seo_page_speed | Measure TTFB, load time, HTML size, and resource counts |
seo_sitemap_parse | Parse sitemap.xml, check freshness, duplicates, and compliance |
Add to your Claude Code MCP configuration:
Add to your claude_desktop_config.json:
Fetches a URL and inspects all meta-related elements. Checks title length (optimal: 30-60 chars), meta description length (optimal: 120-160 chars), canonical URL, viewport tag, language attribute, Open Graph completeness, and Twitter Card setup. Returns a 0-100 score.
Parses all headings on a page and builds a visual hierarchy tree. Flags: missing H1, multiple H1s, skipped heading levels, empty headings, overly long headings (>70 chars), and duplicate headings.
Extracts all anchor tags from a page, resolves relative URLs, deduplicates, and probes each link with a HEAD request (falling back to GET when needed). Classifies links as ok, broken, redirect, timeout, or error. Reports internal vs. external link ratio.
Strips scripts, styles, and navigation, then tokenizes the remaining body text. Computes frequency and density for single words, bigrams, and trigrams (filtering stop words). When a target keyword is provided, checks whether it appears in the title, H1, meta description, and first paragraph.
Measures Time to First Byte and total download time from the server's perspective. Counts external scripts, stylesheets, images, and iframes. Detects render-blocking scripts (missing async/defer), images without alt text or dimensions, missing compression, and absence of lazy loading.
Accepts a sitemap URL or any page URL (automatically appends /sitemap.xml). Falls back to checking robots.txt for a Sitemap directive. Handles both <sitemapindex> and <urlset> formats. Validates URL count limits, lastmod freshness, duplicate entries, protocol consistency, and trailing slash patterns.
MIT