Executive summary
VNZT now looks less like a conventional site and more like a knowledge framework: human-readable pages, AI Layer, Knowledge Graph, API resources, documentation, versions, project status and llms files work together.
After fixes, the site is accessible to search and AI crawlers, has a clean sitemap without JSON/Markdown, JSON-LD coverage for HTML pages and a synchronized entity graph.
Scores by area
| Technical SEO/GEO | 92/100 | Strong | Sitemap URLs return 200, HTTPS, canonical URLs, headers, robots and AI crawler access work. Remaining issue: HTTP to HTTPS is still a 302 at hosting level. |
| Schema / Knowledge Graph | 90/100 | Strong | HTML pages contain JSON-LD, stable @id values are used, graph/entities/relations are synchronized. |
| AI Visibility / llms | 88/100 | Strong | llms.txt and llms-full.txt describe the platform, model, AI resources, interpretation rules and research status. |
| Content / E-E-A-T | 74/100 | Good but needs strengthening | The research frame is clear, but external references, protocols and validation artifacts need further development. |
| Brand / Entity Authority | 58/100 | Needs work | First-party signals are strong, but independent external evidence is still limited. |
| Platform readiness | 82/100 | Good | Google, Gemini, Perplexity, Bing, ChatGPT and Claude can technically access key pages; citation authority is the next step. |
Fixed after audit
| GPTBot, ClaudeBot and bingbot received 403 | Fixed | They now receive 200 on key HTML and JSON pages. |
| Security headers were missing | Fixed | HSTS, CSP, X-Content-Type-Options, X-Frame-Options, Referrer-Policy and Permissions-Policy were added. |
| JSON-LD coverage was sparse | Fixed | All sitemap HTML pages include JSON-LD. |
| Knowledge Graph projections were inconsistent | Fixed | graph.json, entities.json and relations.json are synchronized. |
Remaining gaps
The main gap is external authority. AI systems can read VNZT, but they still need corroborating signals: public profiles, repositories, OSF/Zenodo artifacts, preprints, citations or independent discussion.
The second gap is empirical validation. H1, H2 and H3 are promising because they can fail, but they still require preregistration, sample sizes, controls, analysis plans and published results.