Datasets:

lucioduran
/

ax-web-readiness

rank int64	domain string	url string	overall_score int64	grade_label string	grade_color string	duration_ms int64	llms_txt_score int64	llms_txt_pass int64	llms_txt_warn int64	llms_txt_fail int64	robots_txt_score int64	robots_txt_pass int64	robots_txt_warn int64	robots_txt_fail int64	structured_data_score int64	structured_data_pass int64	structured_data_warn int64	structured_data_fail int64	http_headers_score int64	http_headers_pass int64	http_headers_warn int64	http_headers_fail int64	agent_json_score int64	agent_json_pass int64	agent_json_fail int64	mcp_score int64	mcp_pass int64	mcp_fail int64	security_txt_score int64	security_txt_pass int64	security_txt_warn int64	security_txt_fail int64	meta_tags_score int64	meta_tags_pass int64	meta_tags_warn int64	meta_tags_fail int64	openapi_score int64	openapi_pass int64	openapi_fail int64	has_llms_txt bool	has_robots_txt bool	has_agent_json bool	has_mcp_json bool	has_security_txt bool	has_openapi bool	timestamp string
4	gtld-servers.net	https://gtld-servers.net	0	Poor	red	299	0	0	0	1	0	0	0	1	0	0	0	1	0	0	0	1	0	0	1	0	0	1	0	0	0	1	0	0	0	1	0	0	1	false	false	false	false	false	false	2026-02-27T22:03:28.055Z
1	google.com	https://google.com	27	Poor	red	685	0	0	0	1	60	2	1	1	0	0	0	1	60	0	2	2	0	0	1	0	0	1	100	5	0	0	30	0	5	0	0	0	1	false	true	false	false	true	false	2026-02-27T22:03:28.438Z
2	microsoft.com	https://microsoft.com	31	Poor	red	2,044	0	0	0	1	60	2	1	1	10	1	1	1	70	0	2	1	0	0	1	0	0	1	100	5	0	0	40	1	4	0	0	0	1	false	true	false	false	true	false	2026-02-27T22:03:29.799Z
5	facebook.com	https://facebook.com	39	Poor	red	2,430	65	2	4	0	54	1	5	0	0	0	0	1	85	1	1	0	0	0	1	0	0	1	100	5	0	0	30	0	5	0	0	0	1	true	true	false	false	true	false	2026-02-27T22:03:30.188Z
3	mail.ru	https://mail.ru	32	Poor	red	5,936	65	2	4	0	25	1	3	1	0	0	0	1	75	1	2	0	10	1	1	10	1	1	45	1	1	2	30	0	5	0	10	1	1	true	true	true	true	true	true	2026-02-27T22:03:33.692Z

AX Web Readiness

The first large-scale dataset measuring how ready the web is for AI agents.

This dataset contains AI Agent Experience (AX) audit results for the top 10,000 websites (ranked by Tranco), generated with ax-audit — an open-source tool that evaluates 9 key standards that AI agents rely on to discover, understand, and interact with web services.

Why This Dataset Matters

As AI agents become primary consumers of web content, websites need machine-readable metadata beyond traditional SEO. Standards like llms.txt, agent.json (A2A), MCP protocol, and structured data enable agents to autonomously navigate and use web services. This dataset quantifies adoption across the web's most visited domains.

Methodology

Source: Tranco top 10K domains (research-grade popularity ranking)
Tool: ax-audit v2.1.0
Protocol: HTTPS GET to each domain's root URL
Timeout: 15 seconds per domain
Date: February 2026

Checks & Weights

Check	Weight	What It Measures
LLMs.txt	15%	Presence and quality of `/llms.txt` per the llmstxt.org spec
Robots.txt	15%	AI crawler configuration in `robots.txt`
Structured Data	13%	JSON-LD / schema.org structured data
HTTP Headers	13%	Security headers + RFC 5988 Link headers for AI discovery
Agent Card (A2A)	10%	`/.well-known/agent.json` — Agent-to-Agent protocol
MCP Protocol	10%	`/.well-known/mcp.json` — Model Context Protocol
Security.txt	8%	`/.well-known/security.txt` per RFC 9116
Meta Tags	8%	AI-relevant meta tags and `rel` links in HTML
OpenAPI	8%	`/.well-known/openapi.json` API specification

Schema

Each row represents one audited domain with ~45 flat columns:

Column	Type	Description
`rank`	int	Tranco rank (1 = most popular)
`domain`	string	Domain name (e.g., `google.com`)
`url`	string	Full URL audited
`overall_score`	int	Weighted AX readiness score (0–100)
`grade_label`	string	`Excellent` / `Good` / `Fair` / `Poor`
`grade_color`	string	`green` / `yellow` / `orange` / `red`
`duration_ms`	int	Audit duration in milliseconds
`{check}_score`	int	Individual check score (0–100)
`{check}_pass`	int	Number of passing findings
`{check}_warn`	int	Number of warning findings
`{check}_fail`	int	Number of failing findings
`has_llms_txt`	bool	Whether `/llms.txt` exists
`has_robots_txt`	bool	Whether `/robots.txt` exists
`has_agent_json`	bool	Whether `/.well-known/agent.json` exists
`has_mcp_json`	bool	Whether `/.well-known/mcp.json` exists
`has_security_txt`	bool	Whether `/.well-known/security.txt` exists
`has_openapi`	bool	Whether `/.well-known/openapi.json` exists
`timestamp`	string	ISO 8601 audit timestamp

Check prefixes: llms_txt, robots_txt, structured_data, http_headers, agent_json, mcp, security_txt, meta_tags, openapi.

Reproduce

git clone https://github.com/lucioduran/ax-web-readiness.git
cd ax-web-readiness
npm install
node crawl.js --limit 10000

License

This dataset is released under CC BY 4.0. The underlying tool (ax-audit) is licensed under Apache 2.0.

Citation

@dataset{duran2026axwebreadiness,
  title={AX Web Readiness: AI Agent Experience Audit of Top 10K Websites},
  author={Duran, Lucio},
  year={2026},
  url={https://huggingface.co/datasets/lucioduran/ax-web-readiness},
  note={Generated with ax-audit v2.1.0}
}