AI Agent Readiness Checker
Find out whether AI search, chatbots and agents can find and read your website. The checker fetches your public files and tests crawler access for every major AI bot, llms.txt, server-rendered content, structured data, metadata and the accessibility signals agents rely on, then tells you exactly what to fix.
The checker requests your home page, /robots.txt, /llms.txt, /llms-full.txt and your sitemap, once each, identifying itself as BraceAgentReadiness. Results are not saved. It cannot test pages behind a login.
Which AI bots can fetch your home page
From your robots.txt. A bot with its own group ignores the rules for *. Training bots collect data to build models; search and user-triggered bots fetch pages to answer questions.
What a crawler without JavaScript reads
Need to fix something? Build a robots.txt and llms.txt with AI Crawler Controls, and check your tags with the Index Checker.
What makes a site ready for AI agents
People now reach websites through ChatGPT, Claude, Gemini, Perplexity and agents that browse for them. These tools fetch pages differently from a browser, and most do not run JavaScript, so a site can look perfect to visitors and nearly empty to an AI. Readiness comes down to three questions.
1. Can they get in? (crawl access)
Your robots.txt decides which bots may fetch which pages. Blocking training crawlers (GPTBot, ClaudeBot, Google-Extended) opts your content out of model training, but blocking search and user bots (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot) removes you from AI answers and stops assistants from opening your pages for users. Also check that your pages are not marked noindex or nosnippet.
2. Can they read it? (content)
- Server-rendered text: the words must be in the HTML your server sends, not built later by JavaScript.
- llms.txt: an emerging convention, a Markdown file at
/llms.txtwith a title, a one-line summary and links to your best pages. It is optional and no AI company guarantees it is used, but it is cheap and helps tools that look for it. - Structured data and metadata: JSON-LD, a clear title, description and canonical URL tell machines what each page is.
- Markdown versions: agents read Markdown far more cheaply than HTML.
3. Can they act? (built for agents)
Agents that operate a page often work from its accessibility tree, so semantic landmarks, named links and buttons, labelled form fields and image alt text matter. WebMCP, an experimental web standard, lets a page register tools that in-browser agents can call directly; it is reported here but does not affect the score.
How the score works
Each check is worth a set number of points: pass earns all of them, a warning half, a failure none, and informational items none. The total is shown out of 100. The score is a guide to what is fixable, not a ranking or a promise of AI visibility.
How to use the checker
- Enter your domain and press Check site.
- Read the score, then work through the failed and warning checks, each with a fix.
- Review the AI bot table to confirm which crawlers you are allowing.
- Fix the issues, check again, and share the report with your team.
Why use this checker
- Tests every major AI bot against your robots.txt
- Validates llms.txt, structured data and sitemaps
- Shows exactly what a no-JavaScript crawler sees
- Plain-English fixes with copy-paste examples
Related tools
Frequently Asked Questions
What does the AI agent readiness checker test?
It fetches your home page, robots.txt, llms.txt, llms-full.txt and sitemap and checks: whether AI training, search and user-triggered bots are allowed, whether the page returns usable text without JavaScript, structured data and metadata, whether a Markdown version exists, and accessibility signals such as landmarks, named links and buttons, labelled form fields and image alt text.
Do I need an llms.txt file?
No. llms.txt is an emerging convention rather than a standard that every AI company honours, so a missing file only costs a few points here. It is cheap to add, though: a short Markdown file with a title, a one-line summary and links to your key pages gives tools that look for it a curated map of your site.
Should I block AI crawlers in robots.txt?
That is your decision. Blocking training crawlers such as GPTBot, ClaudeBot and Google-Extended opts your content out of model training. Blocking search and user-triggered bots such as OAI-SearchBot, ChatGPT-User, Claude-SearchBot and PerplexityBot means your pages are less likely to be fetched and cited in AI answers. The checker flags the second kind because it usually costs visibility.
Why does the checker care about JavaScript?
Many AI crawlers fetch the raw HTML and do not run JavaScript. If your content is built in the browser, they may see an empty page even though visitors see everything. The tool shows the exact text a crawler without JavaScript can read from your home page.
Is the check private? What does it store?
The checker's server fetches only public files from the address you enter and returns them to your browser. Nothing is stored or logged. It identifies itself as BraceAgentReadiness, only contacts the site you entered (and its www or https variant), and refuses private or internal addresses.
What is WebMCP?
WebMCP is an experimental web standard proposal that lets a web page register tools, as JavaScript functions with descriptions and schemas, that an in-browser AI agent can call. The checker looks for it in a page's inline scripts and reports it as a bonus; it does not affect your score.