Lumenyl AI Search Intelligence

A dependency-free Node.js web app for benchmarking a website's readiness for frontier AI answer engines. Lumenyl evaluates whether systems such as ChatGPT, Claude, Perplexity, Gemini / Google AI Overviews, and Microsoft Copilot can discover, understand, trust, extract, and cite a page.

Lumenyl checks crawler access, server-rendered content, schema.org depth, answer-ready passages, evidence-backed claims, entity authority, llms.txt, sitemaps, and prioritized AI-search visibility actions.

Why this exists

Modern discovery is shifting from ranked links to AI-generated answers. Lumenyl translates enterprise GEO, prompt intelligence, citation analysis, and entity optimization patterns into an actionable readiness brief that answers: can frontier AI systems discover, understand, trust, extract, and cite this page?

What it analyzes

Lumenyl produces a scored AI readiness report across six weighted dimensions:

• AI crawler access & indexability — robots.txt availability, retrieval-crawler permissions, on-demand fetcher access, training-crawler policy separation, and sitemap discovery.
• Answer quality & citation potential — answer-first summaries, numeric/statistical claims, authoritative citations, concise passages, FAQ-style coverage, expert language, original data, and freshness signals.
• Structured data & entity intelligence — JSON-LD/schema.org coverage, Organization/Product/Person-style entity identifiers, title and meta description quality, heading hierarchy, and Open Graph summaries.
• Technical retrieval performance — HTML content type, server-rendered text availability, response speed, canonical tags, language metadata, and viewport metadata.
• Brand authority & entity footprint — consistent brand signals, sameAs/entity links, authoritative outbound references, author/reviewer cues, and public knowledge/community signals.
• Emerging AI standards — llms.txt, sitemap hygiene, AI/provenance disclosure headers, and early AI content-usage preference signals.

Report output

Each audit returns:

• An overall readiness score and status, from critical blockers to AI citation-ready.
• Section-level scores for access, content, structure, technical, authority, and standards.
• An intelligence snapshot with detected brand, visible content depth, schema types, authority signals, robots.txt, llms.txt, and sitemap status.
• A prioritized AI visibility roadmap sorted by severity, impact, and effort.
• Diagnostic checks with evidence, strategic rationale, recommendations, and next steps.
• A prompt portfolio to monitor in frontier answer engines.
• A benchmark panel mapping Lumenyl's checks to common AI search intelligence platform patterns.

Features

• Runs locally with Node.js 18+ and no runtime dependencies.
• Fetches and analyzes a target page, robots.txt, llms.txt, and sitemap candidates.
• Distinguishes AI retrieval crawlers from training crawlers and user-triggered fetchers.
• Parses visible HTML, metadata, headings, paragraphs, links, JSON-LD, and numeric claims.
• Provides a browser UI plus a JSON API for programmatic audits.
• Uses Node's built-in test runner for analyzer coverage.

Run locally

Requirements:

• Node.js 18 or newer

Install or clone the project, then run:

npm test
npm start

Open http://localhost:3000, enter a URL, and run the AI readiness analysis.

For development with automatic server restarts:

npm run dev

API

The local server exposes a single audit endpoint:

curl -s http://localhost:3000/api/audit \
  -H 'Content-Type: application/json' \
  -d '{"url":"https://example.com"}' | jq

The response includes the analyzed URL, score object, summary, page facts, section diagnostics, prioritized actions, prompt portfolio, and market benchmark metadata.

Project structure

public/             Browser UI assets
  index.html        App shell and audit form
  app.js            Report rendering and client-side API call
  style.css         Visual styling
src/
  server.js         Dependency-free static server and /api/audit endpoint
  analyzer.js       URL normalization, fetching, parsing, scoring, and recommendations
test/
  analyzer.test.js  Node test runner coverage for analyzer behavior

Example use cases

• Check whether a product, documentation, or thought-leadership page can be retrieved and cited by AI answer engines.
• Find accidental robots.txt blocks for AI retrieval crawlers without conflating them with training opt-outs.
• Identify pages that rely too heavily on client-side rendering for citation-worthy content.
• Prioritize content updates that add answer-first summaries, evidence, statistics, FAQs, and entity clarity.
• Track readiness for emerging AI discovery standards such as llms.txt and AI usage preference headers.

Limitations

Lumenyl is a readiness benchmark, not a guarantee of inclusion in any AI answer engine. Frontier systems use private ranking, retrieval, freshness, safety, personalization, and citation-selection logic that can change without notice.

The analyzer inspects public page artifacts from a single URL at audit time. It does not log into private sites, execute full browser rendering, crawl an entire domain, measure actual AI citations, or query external LLM/search APIs.

License

Apache-2.0