Recommendations

Actionable fixes the Analyzer surfaces for each signal category.

How recommendations work

After scoring your site, the Analyzer generates a prioritized list of recommendations — one per missing or malformed signal. Each recommendation includes a description of the issue, the expected fix, and a link to the relevant documentation.

llms.txt recommendations

• Missing llms.txt — no file found at /llms.txt. Generate one with geo-ai-generate or createGeoAI().generateLlms().
• Invalid format — file exists but does not follow the llms.txt standard. Check that it starts with an H1 heading and uses the correct Markdown structure.
• Missing llms-full.txt — the extended file is not present. Pass true to generateLlms() to generate it.
• Empty sections — llms.txt has no resource entries. Add pages, products, or other content to your provider.

AI metadata recommendations

• Missing meta tags — add meta name="llms" and meta name="llms-full" to your page head.
• Missing Link header — add the Link: </llms.txt>; rel="ai-content-index" HTTP header. Use injectLinkHeader: true in the GEO AI middleware.
• Broken meta tag URL — the URL in the meta tag returns a non-200 response. Verify your llms.txt is publicly accessible.

Crawler rules recommendations

• No robots.txt — create a robots.txt file and add explicit allow rules for AI crawlers.
• AI bots blocked — one or more AI crawlers are disallowed. Update your crawler config to allow the bots you want to index your site.
• No explicit AI rules — robots.txt exists but has no AI-specific directives. Add them with geo.generateRobotsTxt().

Structured signals recommendations

• No JSON-LD — no application/ld+json script found. Add Schema.org markup with geo.generateJsonLd().
• Missing required fields — JSON-LD is present but missing name, url, or description. Ensure your provider supplies these fields for all resources.
• Unknown schema type — the @type is not recognized by AI search engines. Use WebSite, Product, or Article.