AXRAY
Sign inCreate account

For agencies and consultants

Hand your client a report with your name on it

Crawl a client's whole site, get every page scored against a published rubric, and download a PDF that says your agency across the top. One scan takes about two seconds and needs nothing installed on their side.

Look at the deliverable first

Everybody in this job asks the same first question, so here is the answer before the pitch. This is the PDF a client receives — a real report of example.com, produced by the same code that produces yours, with a fictional agency's name on it where yours would be.

We report on example.com rather than on ourselves for two reasons: our own site scores 100 and therefore has no fix list, which is a useless demonstration for somebody whose job is handing over a list of fixes; and example.com is reserved by IANA, so no real business is being made an example of.

The job this does

  1. Crawl the client's site. One address in, up to 500 pages on Team and 2000 on Scale. You get the problems that repeat across a template, which are the ones worth an invoice, rather than whatever was wrong with the page you happened to open.
  2. Read the fix list. Every finding carries the points it recovers, why an agent cares, and the exact lines to paste. It is written to be handed to a developer without translation.
  3. Download the PDF under your name. Masthead, footer, done. Attach it, bill it, and re-run it next quarter to show the number moved.

Why a client will listen

Because the numbers are uncomfortable and they are not ours to spin. Of 292 well-known public sites we measured on 2026-09-06:

68/100 median score — half the web reaches an assistant worse than this
72 serve a bot wall to an ordinary agent request instead of their content
11 send almost nothing at all without JavaScript

The whole dataset is open and downloadable under CC BY 4.0, method and limits included, and no site in it is named. You can put those figures in a pitch deck and point at the source, which is not true of most numbers in this subject.

What this is not

Worth saying plainly, because you are the one who has to defend it in a room.

  • It is not an SEO tool and does not replace one. It measures what a reader with no browser receives. Google runs a full browser. The two disagree, and where they disagree is the interesting part.
  • It cannot tell you how much traffic a fix will bring. Nobody can, and a tool selling honest measurement is the last place that number should be invented.
  • The command line is free and unlimited. If you need an occasional look rather than a deliverable, npx axray-cli client-site.com --crawl 50 costs nothing and we would rather you knew.

Questions

Whose name is on the report?

Yours. On Team and above you set a name and an address, and they replace the masthead and the footer. What stays on every copy is the line naming the rubric, its version and where it is published — partly because a report claiming to be your own measurement with nothing saying what measured it is exactly the unverifiable claim this tool exists to argue against, and partly because it is what makes your number worth quoting. Your client can re-run the same scan and get the same figure. That is the difference between an audit and an opinion.

Can I run this on a client site without their involvement?

Yes. Everything is measured from one ordinary HTTP request to a public page — no account on their side, no tag to install, no access to anything. It is the same request an AI crawler makes, which is the point.

Does anything about my client become public?

No. A report is reachable only by its own unguessable link, and a host appears in the public AX Index only when somebody proves they control the domain and asks to be listed. Neither happens by scanning.

How is this different from an SEO audit?

It measures a different reader. An SEO tool asks what Google, running a full browser, makes of a page. This asks what an assistant with no browser and no JavaScript receives — which for a lot of sites is almost nothing, and which no SEO tool reports because it is not their question. Run both; they disagree in useful places.

What if I only need this occasionally?

The command-line version is free and has no limit: npx axray-cli client-site.com --crawl 50 runs the same scanner on your own machine and prints the same findings. It cannot put your name on a PDF, and it has no history to show a client next quarter, but for an occasional look it costs nothing and we would rather say so.

Try it on a client site now. One scan, about two seconds, no account and nothing installed.