# Is your site ready for AI agents? > Free scan of any website for AI agent readiness: llms.txt, Markdown for agents, Link headers, API catalog, agent skills, MCP and more, with a fix for each gap. Canonical page: https://aispeedforce.com/agent-ready/ Last updated: 2026-09-22 AI assistants and agents now read, compare and act on websites for people. This scan checks whether yours gives them what they look for: the files, headers and endpoints that let an agent find you, read you cheaply and use what you offer. ## What the scan checks Six groups of checks. Core checks count double, advanced checks count once, and the authentication and payment groups are reported but never scored. ### Discoverability 5 checks - 3 core, 2 advanced - robots.txt Core The first file any crawler or agent reads. It says what may be fetched and where the sitemap is. - XML sitemap Core Lists every page you want found, so an agent does not have to guess your URLs. - llms.txt Core A short Markdown guide for language models: what the site is and where the important pages are. - Link response headers Advanced HTTP Link headers on the homepage point agents straight at your machine-readable resources. - DNS-AID records Advanced Lets agents find your agent endpoints through DNS, before fetching anything from your site. ### Agent discovery and APIs 6 checks - 6 advanced - API catalog Advanced One well-known file listing the APIs you offer and where their descriptions are. - Agent Skills index Advanced Publishes ready-made instructions (SKILL.md files) that teach an agent how to work with your site. - AI catalog (ARD) Advanced A manifest of everything agentic you offer, from APIs to MCP servers, for registries and agents to index. - MCP server card Advanced If you run an MCP server, the card lets agents find it and learn its transport and capabilities. - A2A agent card Advanced If you run an agent other agents can talk to, its card describes its skills and endpoint. - WebMCP tools Advanced Exposes your site's actions as tools an AI agent in the visitor's browser can call directly. ### Crawler and AI bot rules 2 checks - 2 core - AI crawler rules Core Naming AI crawlers in robots.txt makes your policy explicit instead of inherited from a wildcard. - Content Signals Core States whether your content may be used for search, for answering questions, and for model training. ### Content for agents 3 checks - 3 core - Markdown on request Core Agents read Markdown far more cheaply than HTML. Asking for it with Accept: text/markdown should return it. - Markdown alternate link Core A link in the page head tells an agent where the Markdown copy of that page lives. - Structured data (JSON-LD) Core Schema.org data states facts about the page and the business in a form agents do not have to infer. ### Authentication 3 checks - not scored. Only relevant if agents need to log in to an API you run. - OAuth / OpenID discovery Not scored Tells agents how to obtain tokens for your protected APIs. - Protected resource metadata Not scored Tells an agent which authorization server issues tokens for a given API. - auth.md Not scored Plain-language instructions for how an agent registers and authenticates. ### Agent payments 3 checks - not scored. Only relevant if agents can buy from you directly. - UCP profile Not scored Universal Commerce Protocol profile for agents that buy on a user's behalf. - ACP discovery Not scored Agentic Commerce Protocol discovery for agent checkouts. - Machine payments (MPP) Not scored Declares payable API operations so agents can pay per call. ## How to read the score The score is out of 100. A pass earns full weight, a partial pass earns half, and a check that could not be completed is left out rather than counted against you. - **90 to 100**: Agent-native. Agents can find, read and use the site with no guesswork. - **70 to 89**: Agent-ready. The basics are in place and most of the discovery layer is too. - **40 to 69**: Partly ready. Agents can read the site, but they have to work harder than they should. - **0 to 39**: Not ready yet. Agents fall back to scraping HTML and guessing, and are more likely to get you wrong. Not every check applies to every site. An MCP server card only makes sense if you run an MCP server. The report says which gaps are worth closing and which you can leave. ## About the scan ### Is the scan free? Yes. There is no sign-up and no email address needed. Enter an address and the report appears on this page. ### What do you store when I run a scan? The report for each site is kept for ten minutes so a repeat scan is instant, and a hashed form of your IP address is kept for rate limiting. Both are deleted within an hour. The privacy policy has the details. ### Why are authentication and payment checks not scored? They only matter if agents need to log in to an API you run, or buy from you directly. A site without either should not lose points for leaving them out, so they are listed for completeness and left out of the score. ### Does the scan run the page's JavaScript? No. It reads the page and its own scripts as text. WebMCP is detected by finding the navigator.modelContext call in that code, not by running it, and the report says so. ### Can an AI agent use this scanner? Yes. Request /agent-ready/scan.php?format=json&url=example.com and the full report comes back as JSON. The same rate limits apply. ### Can you fix the gaps for us? Yes, as a scoped project like any other. Send the report and the site to us through the contact page and we will say which gaps are worth closing for your site and which do not apply. ## Want the gaps closed? We build the agent layer as part of our web and agent work: llms.txt, Markdown for agents, API catalogs, skills and WebMCP tools, scoped to what your site actually offers.