AgentLighthouse

Free AI Readiness & Modern Web Compliance Scanner

Scan your site live · API Docs · MIT License

Is your website ready for AI agents and modern web standards? Find out in 10 seconds.

---

Why Agent Readiness & Modern Web Compliance Matter

AI agents from Google, OpenAI, Anthropic, and Perplexity are actively crawling the web. If your site lacks structured data, semantic HTML, or llms.txt, these agents can't understand your content properly.

Meanwhile, Google's Modern Web Guidance defines the standards for modern, accessible, and performant websites — native browser APIs over JavaScript libraries, semantic HTML over divs, and user preference respect.

AgentLighthouse scans any URL and gives you a 0-100 score across 40 checks in 6 categories:

Category	Weight	What It Checks
Security	18pts	SSL, HSTS, CSP, X-Frame-Options, permissions
SEO & Metadata	8pts	Title, description, OG tags, Twitter Card
AI & Agent Readiness	21pts	llms.txt, semantic HTML, AI bot access, headings
Structured Data	14pts	JSON-LD, Schema.org coverage
Modern Web	20pts	Dialog, Popover API, Container Queries, View Transitions, user preference queries, landmarks
Performance & Crawling	19pts	Viewport, charset, canonical, robots.txt, sitemap, compression

Score Levels

• Excellent — 90+ (your site is fully AI-ready and modern)
• Good — 70-89 (minor fixes recommended)
• Fair — 50-69 (important gaps found)
• Poor — Below 50 (significant work needed)

Try It Now

agentlighthouse.burakgider.com

No signup, no API key. Paste a URL, get your score.

API

For CI/CD integration or batch scanning:

curl -X POST https://agentlighthouse.burakgider.com/api/audit \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'

Returns a JSON object with score, level, checks[] array, and per-category breakdown.

AI-Friendly Endpoints

GET /llms.txt        — Machine-readable docs for AI agents
GET /llms-full.txt   — Detailed documentation (Accept: text/markdown)

Modern Web Compliance Checks

Based on Google Chrome Modern Web Guidance:

• Native Dialog — <dialog> element instead of custom modals
• Responsive Images — <picture> + srcset for adaptive delivery
• Container Queries — @container for component-level responsive design
• Popover API — Native popover attribute over tooltip libraries
• View Transitions — Smooth page transitions API
• User Preference Queries — prefers-reduced-motion + prefers-color-scheme
• Landmark Elements — <main>, <nav>, <header>, <footer>, <aside>, <article>
• Resource Hints — preload, prefetch, preconnect
• Time Element — Machine-readable <time datetime="...">
• Native Disclosure — <details>/<summary> over custom accordions
• Inline Event Handlers — CSP-compatible addEventListener over inline handlers
• CSP Meta Tag — Fallback Content-Security-Policy in <meta>

Built For the Agentic Era

AgentLighthouse is designed for a world where AI agents discover, read, and act on web content:

• Detects llms.txt / llms-full.txt (the emerging standard for AI-readable docs)
• Validates AI crawler access rules in robots.txt (GPTBot, Claude-Web, PerplexityBot)
• Checks semantic HTML structure that agents rely on for comprehension
• Analyzes JSON-LD / Schema.org structured data coverage
• Scans for Google Modern Web Guidance compliance

Tech Stack

Python 3.11, Flask, BeautifulSoup4, lxml, vanilla JS

License

MIT · PHRIXUS

---

Self-Host / Developer Setup

git clone https://github.com/phrixus-ai/agent-lighthouse.git
cd agent-lighthouse
python -m venv .venv && source .venv/bin/activate
pip install -e .
cp .env.example .env  # optional — for production
python -m webaudit     # → http://localhost:5000

Variable	Default	Description
WEBAUDIT_PORT	5000	Flask port
SITE_URL	(empty)	Production URL for SEO/OG tags
AGENTLIGHTHOUSE_SECRET_KEY	dev-key-change-in-production	Flask secret key