Agent Readiness Checker
by
Score your domain on the files, formats and interfaces AI agents actually look for.
We probe the files, headers and well-known endpoints that AI agents look for, then score your domain across four dimensions and tell you what to fix first.
Discovery files
Machine-readable content
Crawler access rules
Agent interfaces
What is agent readiness?
AI agents look for a specific set of files, headers and well-known endpoints before they read a single page. This check tells you which of them you publish, which you are missing, and which of your rules are quietly blocking the crawlers you want.
Free Agent Readiness Check
We check public files on your domain only. No login or code change required.
Why agent readiness matters
AI agents do not browse your site the way a person does. They look for a small set of files at known paths, prefer content in formats they can parse without a renderer, and read your robots rules to decide what they are allowed to do. When those pieces are missing, an agent falls back to guessing, and guessing loses you citations.
This checker inspects the same signals an agent would, scores them across four dimensions, and hands you the list of gaps. It is free, requires no access to your site, and finishes in under a minute.
How the check works
Probe
We probe your site root for the files and headers agents look for, using read-only requests.
Score
Each signal is scored inside its dimension, weighted by how much it affects real agent access today.
Fix
Failed checks become a prioritised list, each one pointing at the specification that defines the fix.
What we check
Each check maps to a published specification or a documented convention, so a fix is always concrete.
robots.txt and crawler rules
The file every crawler reads first. We confirm it is reachable and parse the rules that apply to each AI crawler by name.
llms.txt
A curated, model-facing reading list at your root. We check it exists and that it returns plain text rather than an HTML shell.
Markdown negotiation
We ask your homepage for text/markdown and record what comes back, then look for a .md twin as the simpler alternative.
Content Signals
A robots.txt directive that separates search indexing, AI inference, and AI training, so you can allow one without allowing all three.
MCP server card
A JSON document describing the tools your service exposes to AI clients, and the transport they should use to reach them.
API catalog and OpenAPI
One well-known document listing your public APIs, so an agent can call you instead of scraping you.
What we score
The four dimensions of agent readiness
Four scored dimensions, plus an informational look at the commerce standards taking shape.
Discovery
robots.txt, your XML sitemap, llms.txt and HTTP Link headers. The files an agent reads before it reads anything else.
Machine-readable content
Markdown content negotiation, markdown URL variants, JSON-LD and feed autodiscovery. Content a model can use as-is.
Access and permissions
Per-crawler robots rules, the Content Signals directive, a live fetch with a crawler user agent, and Web Bot Auth.
Agent interfaces
MCP server card, agent skills index, API catalog, OpenAPI, OAuth discovery, A2A and ARD catalogs.
How it works
How the agent readiness check works
One domain in, a scored report out.
Enter your domain
Enter your domain. We only need the root, no access or code change.
We probe the signals
We run about twenty read-only requests against your site root, covering discovery files, content negotiation, crawler rules and well-known agent endpoints.
Read your report
You get a score per dimension, every individual check with its evidence, and a prioritised list of fixes.
Frequently asked questions
What is agent readiness?
Agent readiness is how easy your site makes it for an AI agent to find, read, and act on what you publish. It covers the files agents look for, the formats they prefer, the rules you set for them, and the interfaces you expose. It is separate from how often models mention you, which is what visibility tracking measures.
How do you test whether crawlers are blocked?
We request your homepage with a real crawler user agent and record what your edge returns. Because the request comes from our network rather than a crawler's published IP ranges, a block here means the user agent string is being refused. A site that verifies crawlers by IP or signature may still block us while allowing the real bot.
Does the check crawl my whole site?
No. Everything we do is a read-only HTTP request to your site root, roughly twenty of them. We do not crawl your pages, submit forms, or use a browser.
Why is my interfaces score so low?
Most sites score low on agent interfaces today, because MCP server cards and API catalogs are new. That is the point. The dimension is weighted so the basics still carry the score, and the interface checks show you where the ceiling is.
Agent readiness questions
Common questions about agent readiness and how the checker works.
Explore more free AI search tools
LLM Pulse is the all-in-one AI search and GEO platform. Use these free tools to audit how AI engines see your brand, then track it all in one place.
See how often AI engines mention your brand across ChatGPT, Perplexity, and Gemini.
Check whether AI answers mention your brand for the prompts that matter.
Grade how well a page is optimized for AI answer engines.
Generate an llms.txt file so AI crawlers understand your site.
Check whether AI crawlers are allowed to read your site.
Check whether your site appears in Common Crawl, the open corpus that trains most AI models.
Test whether AI engines can crawl and render your pages.
Audit your structured data so AI and search engines understand your pages.
See if your content is ready to be cited in AI answers.
Measure how discoverable your brand is across AI engines.
Check whether your site structure helps AI engines understand you.
Validate your product feed for ChatGPT shopping results.
See how AI models answer questions about your brand
Agent readiness tells you whether models can reach you. LLM Pulse tells you whether they actually cite you.