§9

Full audit details

AI assistants can find, read and act on this site.

https://shopworthy-03696d0babc5.herokuapp.com/

Reach100

Can agents get in?

Understand100

Can agents tell what you offer?

Act100

Can agents take the next step?

Trust100

Can agents verify the deal?

Where a typical agent gets stuck

  1. Get inAllowed in and served the real pagePass
  2. Read the pageContent readable without JavaScriptPass
  3. Know who you areKnows your name, logo and contactPass
  4. Find offer & priceGets products with price and stockPass
  5. Take actionFinds a real path to buy, book or contactPass
  6. Check the termsCan confirm returns, shipping and termsPass
All 24 checks: 0 failing, 0 need work, 18 passing

Reach: Can agents get in?100/100

AI assistants allowed in robots.txt Pass

Assistant and AI-search agents from OpenAI, Anthropic, Perplexity, Meta and DuckDuckGo are allowed.

Google and Bing can crawl you Pass

Googlebot and bingbot can crawl your homepage and the 7 key pages we checked. Google AI Overviews and Gemini answer from Google's index, and Copilot from Bing's.

Agents get the same page as browsers Pass

We requested your homepage as each of the 7 assistant agents (ChatGPT-User, OAI-SearchBot, Claude-User, Claude-SearchBot, PerplexityBot, Perplexity-User and Google-Agent). Every one got the same page a browser gets. This is a likely result, not a verified one: real assistant traffic comes from each vendor's published IP ranges, and our test request does not, so a CDN may treat the two differently. Confirm in your CDN settings or server logs.

Content Signals allow AI answers Pass

robots.txt declares "Content-Signal: search=yes, ai-input=yes, ai-train=yes" for ChatGPT-User, OAI-SearchBot, Claude-User, Claude-SearchBot, PerplexityBot, Perplexity-User, Meta-ExternalFetcher, DuckAssistBot, Googlebot, bingbot and all agents. Search and AI answers are allowed.

Pages can be indexed Pass

Your homepage and the 8 other pages we crawled can be indexed: no noindex, and canonicals point to this site.

Content readable without JavaScript Pass

9,092 characters of text are readable without JavaScript.

Sitemap available Pass

Search engines and AI crawlers can discover your full set of pages.

llms.txt (optional) Info

Found /llms.txt. It's optional: there's little evidence AI assistants use it today (Ahrefs found 97% of llms.txt files got no requests at all in May 2026; source: https://ahrefs.com/blog/llmstxt-study/), and Google says no special AI files are needed. Some coding agents do read it, so it's cheap to add but it isn't scored.

Model-training crawlers Info

Training crawlers are allowed. You can block GPTBot, ClaudeBot and CCBot without hurting assistant visibility. Google-Extended is different: Google says it also covers grounding in the Gemini app and API, so blocking it can keep you out of some Gemini answers.

Pages answer without rate limits Pass

Every page we requested answered without a rate limit.

Understand: Can agents tell what you offer?100/100

Descriptive title and summary Pass

Title: "Shopworthy: are you on the AI shelf?". Meta description present.

Structured data (schema.org JSON-LD) Pass

Found Offer ×10, UnitPriceSpecification ×7, Article ×7, Organization ×2, ContactPoint ×2, SoftwareApplication ×2, WebSite and QuantitativeValue across 3 pages.

Business identity is machine-readable Pass

Organization "Shopworthy" with name, url, logo and contact details.

Products and prices are machine-readable Pass

1 offering marked up ("Shopworthy"). Each has a price, currency and availability.

Structured prices match the page Info

We didn't find product pages with product markup to compare against what shoppers see.

Semantic page structure Pass

H1: "When shoppers ask ChatGPT what to buy, does it recommend…" · <main> landmark present · no images.

Act: Can agents take the next step?100/100

Clear, followable paths to act Pass

Found 10 real action paths ("Pricing", "Free", "Pro" and "Growth").

Form fields are labeled Pass

12 of 12 form fields have a proper label. Agents filling a form rely on labels, not placeholder text, to know what goes where.

Direct interface for agents (MCP, A2A, OpenAPI) Info

No MCP server card, A2A agent card or OpenAPI spec at the usual paths. These are emerging standards (the MCP card path is still a draft), and assistants don't use them to decide whom to recommend today, so this isn't scored.

Agent checkout profile (UCP) Info

No /.well-known/ucp. That only matters if you sell products online.

Contact details are machine-readable Pass

Contact details in structured data and 1 tel:/mailto: link.

Trust: Can agents verify the deal?100/100

Policies an agent can check Pass

Linked: terms and privacy. A careful agent checks your terms and privacy policy before it signs someone up or shares their details.

Ratings and reviews in structured data Info

No aggregateRating or Review markup. Agents weigh reviews when choosing between options.

Served over HTTPS Pass

Final URL is HTTPS.

robots.txt rules and live requests
Agentrobots.txtLive request
Assistants and AI search: these should get in
ChatGPT-UserOpenAI · Fetches pages when a ChatGPT user asksAllowedServed HTTP 200
OAI-SearchBotOpenAI · Indexes for ChatGPT searchAllowedServed HTTP 200
Claude-UserAnthropic · Fetches pages when a Claude user asksAllowedServed HTTP 200
Claude-SearchBotAnthropic · Indexes for Claude searchAllowedServed HTTP 200
PerplexityBotPerplexity · Indexes for Perplexity answersAllowedServed HTTP 200
Perplexity-UserPerplexity · Fetches pages when a Perplexity user asksAllowedServed HTTP 200
Google-AgentGoogle · Browses and acts on sites for Gemini users (agent mode)AllowedServed HTTP 200
GooglebotGoogle · Indexes for Google Search, AI Overviews, AI Mode and Gemini groundingAllowednot tested
bingbotMicrosoft · Indexes for Bing and Copilot (and is widely reported to feed ChatGPT search)Allowednot tested
Meta-ExternalFetcherMeta · Fetches pages when a Meta AI user asksAllowednot tested
DuckAssistBotDuckDuckGo · Fetches pages for DuckDuckGo AI answersAllowednot tested
Model training: your call
GPTBotOpenAI · Collects model training dataAllowednot tested
ClaudeBotAnthropic · Collects model training dataAllowednot tested
Google-ExtendedGoogle · Controls Gemini training use (and grounding in the Gemini app and API)Allowednot tested
Applebot-ExtendedApple · Controls Apple AI training useAllowednot tested
CCBotCommon Crawl · Open crawl widely used for trainingAllowednot tested
meta-externalagentMeta · Collects model training dataAllowednot tested

Live requests use representative user-agent strings, so results are "likely". Confirm in your CDN settings or server logs.

The agent's-eye view of your homepage
34.1 KBHTML
9,092characters of readable text
3scripts
Skip to content
Shopworthy
AI shopping Guides Pricing Agencies Method Free tools
Sign in Check my store
AI shopping Guides Pricing Agencies Method Free tools Data study Sign in
Check my store
For online stores · 4 AI assistants
When shoppers ask ChatGPT what to buy, does it recommend you?
Check your store free. See what AI shopping assistants can read on your site, which brands they name instead of you, and the one fix to ship first.
Your store's website
Check my store
More options
Competitors (up to 3: a website or a brand name)
Competitor 1
Competitor 2
Competitor 3
Category
Work it out from my site Coffee Tea Clothing and apparel Activewear Shoes and footwear Outdoor gear Skincare Makeup and cosmetics Hair care Fragrance and perfume Supplements and vitamins Pet supplies Home decor Candles and home fragrance Furniture Mattresses Bedding and bath Kitchenware and cookware Cleaning and household Jewelry Bags and accessories Eyewear Luggage and travel bags Baby and kids Fitness equipment Consumer electronics Chocolate and sweets Food and snacks Drinks and beverages Wine, beer and spirits Toys and games Books and stationery Garden and plants Bikes and cycling Sporting goods Other online store Software and apps News and media Services Buyer questions are built from this category.
Answers
Live when a key is set, otherwise sample Sample only Live answers are on for Claude. Each…
9 pages checked
URLHTTPTextStructured data
/2009,092Organization, ContactPoint, WebSite, SoftwareApplication, Offer, UnitPriceSpecification
/shopping2003,959none
/pricing2003,972Organization, ContactPoint, SoftwareApplication, Offer, UnitPriceSpecification, QuantitativeValue
/shopify2002,533none
/tools/feed2001,442none
/privacy2007,391none
/terms2003,759none
/tools/checkout2002,209none
/guides2003,036CollectionPage, Article