For business owners
How to check if AI agents can use your website
Fetch your homepage without running JavaScript and look at what is actually there. If your prices, services, opening hours or booking links only appear after scripts run, an AI agent cannot see them — and that is the single most common reason a site that looks perfect to a person is unusable to an agent.
You can check this yourself in about a minute, or run a free scan that does the same thing and names each gap. Below is how the check works, what it typically finds, and what to fix first.
The 60-second version you can do yourself
Open your site in a browser with JavaScript turned off. What you see is roughly what an agent sees when it fetches your page. If it looks broken, empty, or like a navigation menu with nothing behind it, that is the problem — and it is invisible to you, because your browser runs the scripts.
In Chrome you can approximate this by disabling JavaScript in the site settings for your domain. The screen that results is not pretty. That is the point.
What agents actually need from a business site
The list depends on what kind of business you are, which is why a generic checklist is nearly useless. A pizza shop needs to expose its sizes, toppings, prices, delivery fee and ordering. An online store needs its products and a way to buy. A plumber needs services, coverage area and a way to book. A software company needs plans and a signup path.
Across all of them, four things come up again and again:
- What you offer — services, products, or menu, as text an agent can read.
- What it costs — prices, or at least a range and how to get a quote.
- When you are available — hours, delivery windows, lead times.
- How to start — a booking, order, or contact route an agent can follow.
Most sites expose the last one as a phone number, which is easy. The first three are usually locked inside scripts or a form, which is where agents give up.
Why “just add a file” is not the answer
There is a lot of advice online that you need a particular file, tag, or plugin to be “AI ready”. Two things genuinely help, and neither is a switch you flip:
Structured data. JSON-LD in your page describing your business, products, prices and hours. Search engines already read products, prices and hours. Search engines already read this, and agents read it too.
A machine-readable summary. A plain text file such as llms.txt that states what the business does and links to the important pages. It costs almost nothing to add and it gives an agent a starting point that is not a guess.
Both of those only help if the facts underneath are real and readable. If your prices are rendered by a script, no file you add will hand them over.
How the free scan checks it
The scan fetches your public homepage the way an agent would — no JavaScript, no browser — and reads what comes back. It then works out what kind of business you are from the evidence on the page, and checks you against what an agent needs from that specific kind of business.
Two honest limits, because they matter more than the pitch:
- If we cannot read your page, we say so. Sites that build everything in the browser give us a shell. Rather than guess at what you are missing, we tell you we could not see it.
- If we cannot tell what kind of business you are, we say that too. A generic checklist is not worth much, so we do not issue a verdict against one. We tell you what you do and check you against the right list.
- Some sites block us. Large sites run bot protection, and a scanning tool looks like a scraper. When that happens we say we were refused, not that your address is wrong.
What to fix first
In practice, the gaps that block agents hardest are the same ones every time, and they are worth fixing in this order:
- Put your offer in the HTML. Services, products, menu — as text, not only as images or script output.
- Put your prices in the HTML, or state clearly how a price is obtained. An agent that cannot find a price cannot compare you to anyone.
- Add structured data for your business, products and hours. This is the highest-leverage change for the least work, because it is a standard format.
- Make the start of a transaction reachable — a booking URL, an order route, or an email, as a real link.
None of this is a redesign. It is mostly making sure the facts you already publish are in a form that survives being read without a browser.
A score is not a certificate
Several tools will give your site a number for free, and that is genuinely useful as a starting point. A certificate is a different claim: it is checked against what your kind of business specifically needs, verified on your live page rather than promised, re-checked weekly, and it lapses when it stops being true.
Certification is free while we are early — no card, no trial. The only thing we ask is that you are able to edit your website's <head>, or that whoever built it is.
Common questions
How do I check if AI agents can use my website?
Fetch your homepage the way an automated agent does — without running JavaScript — and look at what is actually in that HTML. If your prices, services, hours, availability or booking links are only visible after scripts run, an agent cannot see them. You can do this manually with a browser JavaScript disabled, or run a free scan at mcphubz.com/scan which does the same check and names each gap.
Can AI agents read my website right now?
They can read the raw HTML, which is usually only part of your page. Most modern sites build the visible content in the browser with JavaScript, and agents that fetch the page see navigation and little else. That is the most common reason a site is unusable by an agent even though it looks perfect to a person.
What is the most common thing AI agents cannot get from a website?
Structured facts about the business. An agent can usually find a phone number, because that is plain HTML. It struggles with what the business actually offers, what things cost, when it is open, and how to start a booking or an order — because those are typically rendered by scripts or hidden behind a form.
Does my site need a special file or tag for AI agents?
There is no single required file. Two things help most: structured data in the page (JSON-LD, which search engines already use and agents read too), and a machine-readable summary such as llms.txt describing what the site offers. Neither is a magic switch — the underlying facts have to be present in the first place.
Why do some sites refuse to be scanned?
Large sites often run bot protection that blocks automated readers to stop scrapers, and a scanning tool looks exactly like a scraper. When that happens we say so plainly rather than reporting a score, because we genuinely could not see the page. It is a real limit on automated checking, and it affects the biggest sites most.
Is a high readiness score the same as being certified?
No. A score is a number, and plenty of tools will give you one for free. A certificate is different: it is checked against what your specific kind of business needs to expose, verified on your live page rather than promised, re-checked weekly, and it lapses when it stops being true.
Find out what agents cannot get from your site
It takes about a minute, needs no signup, and names every gap with the fix for each one. If we cannot read your site, we will tell you that instead of inventing a score.