What does Nodex evaluate?

Nodex evaluates crawler access, AI bot policy, structured data, metadata, sitemap quality, markdown access, llms.txt, ai.txt, agent-card, agent-skills, visible content, and implementation readiness.

Does Nodex invent authority claims?

No. Nodex marks legal identity, founder details, public profiles, customer proof, awards, reviews, and media mentions as manual authority inputs unless they are already verifiable.

How should generated files be used?

Generated outputs should be deployed, merged, or reviewed based on their bucket: deploy-ready, merge-ready, manual-required, or backend-required.

Methodology

How Nodex evaluates GEO and Agent-Ready signals

Nodex scans a public website from the perspective of AI assistants, AI search systems, and developer teams that need implementation-ready files. The scan separates machine-readable discovery, visible content quality, structured data strength, brand authority, and backend protocol readiness.

GEO visibility

Checks whether the site gives AI systems clear, citable, crawlable, and structured information about the business, products, support paths, policies, and key pages.

Agent readiness

Checks whether agents can discover service metadata, agent-card skills, markdown context, AI policy files, sitemap entries, and implementation instructions.

Generated files

Produces deploy-ready files when safe, merge-ready changes when page code needs editing, manual-required authority tasks, and backend-required protocol work.

Signals Nodex checks

Crawler policy, AI bot access, canonical metadata, Open Graph and Twitter metadata, structured data, FAQ coverage, sitemap freshness, llms.txt, llms-full.txt, ai.txt, markdown access, agent-card, agent-skills, contact paths, legal pages, pricing clarity, and page-specific public facts.

What Nodex will not fake

Nodex does not invent founder names, company registration details, customer logos, reviews, awards, media mentions, public profile links, MCP endpoints, UCP endpoints, API capabilities, or commerce actions that are not actually available.

How to validate a deployment

After changes go live, verify robots.txt, llms.txt, llms-full.txt, ai.txt, agent-card.json, agent-skills index, sitemap.xml, page schema JSON-LD, canonical tags, and markdown negotiation, then run a new Nodex scan.

Last updated

This methodology page was last updated on May 6, 2026. Scoring can still be limited by external platform policy, especially managed crawler blocks at CDN or hosting level.