AgentReadyRun a Scan

Agent readiness assessment·Check library 2026-08-23

Every scanner hands you a score.We hand you the files.

AgentReady measures whether AI agents can find, understand and act on your site. Then it writes your llms.txt and your JSON-LD from your own crawled pages, carrying your products, your prices and your contact points.

Free, no signup. We read your homepage and show the Score, the four Category scores and the 3 highest priority Issues.

See what we checkRead the prices
Deep Scan
up to 50 pages
Report
$19, one-time
Rescans
free for 30 days
Example Scan2026-08-23 14:22 UTC

acme.example

41/100Agent
Ready
Score
  • Discoverability34/100
  • Machine Readability28/100
  • Interaction Readiness44/100
  • Agent-Friendliness58/100

11 Issuesfound on 1 page · 3 shown in full

A technical readiness assessment based on automated checks. Not a certification.

01 / The Scan

You watch the requests happen. That is why the number is believable.

No spinner and no progress bar. A monospaced log writes itself, one line per Check, with the state we recorded.

$ agentready scan acme.example

Free · 1 page

  1. GEThttps://acme.example/

    200 · 1.2 s · body empty until JavaScript runs

    Warning

  2. GEThttps://acme.example/robots.txt

    200 · present · no rule for any named agent

    Warning

  3. GEThttps://acme.example/llms.txt

    404 · absent

    Fail

  4. GEThttps://acme.example/sitemap.xml

    200 · 412 URLs · last modified 2026-08-11

    Pass

  5. PARSEJSON-LD blocks

    0 blocks found in <head> or <body>

    Fail

  6. RENDERhttps://acme.example/pricing

    price string found only inside <img alt="">

    Fail

  7. FORMhttps://acme.example/contact

    4 inputs · 0 <label> · 0 autocomplete tokens

    Fail

  8. GEThttps://acme.example/.well-known/mcp.json

    404 · absent

    Warning

SCORE41/100

11 Issues on 1 page · the 3 highest priority are shown in full, never blurred

Technical schematic: a web page decomposed into a grid of blocks, with request and response arrows crossing to nested bracket structures.
Fig. 1 — One page, decomposed into the requests and structures an agent has to get through.

02 / The gap

What a person sees, and what an agent receives.

The site looks finished. The offer, the price and the booking action are not present as data, so an agent arrives and finds nothing it can use.

What a person sees

Acme Studio

Retainer, monthly

$49

Book a call

Legible, well made, and priced. A human reads it in 3 seconds.

What an agent receives

6 Fail
name
"Acme Studio"
url
https://acme.example
offers
[]
price
null
← rendered inside an image
priceCurrency
null
contactPoint
null
openingHours
null
actions
0 discoverable
llms.txt
absent

Same page, parsed the way an agent has to parse it. The price the human read is not a number anywhere in the markup.

03 / The Fix Pack

Not advice. Files, generated from your own pages.

After a Deep Scan you download a complete llms.txt written from the real page inventory, and JSON-LD blocks already populated with your data.

llms.txt

Generated
# Acme Studio
> Brand and web design studio in Lisbon.
> Project work from $4,800. Retainer from $49 per month.

## Services
- [Brand identity](https://acme.example/services/brand-identity): 4-week sprint, from $4,800
- [Web design](https://acme.example/services/web-design): from $9,200
- [Retainer](https://acme.example/pricing): $49 per month, billed monthly

## Contact
- [Book a call](https://acme.example/contact): form, 4 fields, 30 minutes
- studio@acme.example

Serve this at https://acme.example/llms.txt

service-offer.jsonld

Generated
{
  "@context": "https://schema.org",
  "@type": "Service",
  "name": "Brand identity sprint",
  "provider": {
    "@type": "Organization",
    "name": "Acme Studio",
    "url": "https://acme.example"
  },
  "offers": {
    "@type": "Offer",
    "price": "4800",
    "priceCurrency": "USD",
    "availability": "https://schema.org/InStock"
  }
}

Paste this into the <head> of https://acme.example/services/brand-identity

Technical schematic: a file tree of generated documents with brace structures beside them.

Fig. 2 — The Fix Pack, one file per gap we measured.

Where a fact is missing we say so and leave the field blank. We never invent a price.

Ship the pack, then Rescan the same site at no cost for 30 days and watch the number move.

Rescan, after the fix

91/100, was 41/100 on 2026-08-23

04 / How it holds up

The score is reproducible, or it is worthless.

a

Deterministic by construction

Every score comes from HTTP requests, HTML parsing and headless browser checks against explicit rules. The same unchanged site scores the same tomorrow.

b

Evidence, not opinion

Every Fail records what we requested and what came back, with the exact URL, in full. An agency can attach it to a fix invoice.

c

Models draft files, never scores

Language models write the Fix Pack from crawled facts and nothing else. They never produce or adjust a number.

05 / The check library

Four Categories, in this order, forever.

Agent standards land monthly. Each one becomes a new Check, a new Fix, and a reason to Rescan a site you already unlocked.

01

Discoverability

robots.txt rules for named agents, sitemap coverage, llms.txt, Markdown content negotiation, HTTP status of commercial pages.

Example34/100

02

Machine Readability

JSON-LD presence and validity, Offer price and currency as data, heading semantics, content reachable without rendering.

Example28/100

03

Interaction Readiness

labelled form fields, autocomplete tokens, reachable contact, booking and checkout actions, errors an agent can read.

Example44/100

04

Agent-Friendliness

bot access control, challenge pages served to identified agents, rate limits, MCP Server Card, API catalog.

Example58/100

Example scores from the Scan of acme.example shown above. A technical readiness assessment based on automated checks. Not a certification.

06 / Prices

One-time payments. No subscription.

A human audit of the same ground is four figures and takes weeks. This is the same artifact, priced as software.

Scan

Free

Homepage only. Agent Ready Score, 4 Category scores, the 3 worst Issues in full. No signup.

Run a Scan

Report

$19

Deep Scan of up to 50 pages. Per-Check Evidence, priority-ordered correction list, page-by-page results, Fix Pack. Free Rescans for 30 days.

Buy the Report

5-site pack

$79

5 site credits. One credit unlocks one site, whatever it costs us to crawl it.

Buy 5 credits

20-site pack

$199

20 site credits. Built for the agency reviewing a portfolio in one sitting.

Buy 20 credits
  • Site credits never expire.
  • One credit per site, not per scan.
  • Rescans of an unlocked site are free for 30 days from purchase.
  • Purchases are tied to the email you pay with, so a Report is always recoverable.

One field. Then the log.

No signup, no options, no captcha. Enter a site and watch the Checks run. The Scan is free and the result page is permanent, so you can send it to a client as it is.

agentready-hyo.nanocorp.app

Run a Scan

Free

Homepage only, 10 to 30 s. The Deep Scan reads up to 50 pages and writes the Fix Pack, for $19.