Zeodyn Scanner is now on ChatGPT

Sixty per cent of e-commerce sites score “Not Ready” for AI agent commerce. Now you can find out where any site stands without leaving ChatGPT.

Zeodyn Scanner™ is available as a Custom GPT on the GPT Store — the first agent commerce readiness scanner on any AI marketplace. Open it, type a URL, and get a Zeodyn Score™ from 1 to 100 across six dimensions of readiness. No account required.

Try it on ChatGPT →

What it does

The scanner runs the same Agent Commerce Stack™ methodology whether you use it on the web, through the API, or inside ChatGPT. Fifty-four automated sub-checks assess a site across six dimensions: Discovery & Access, Structured Data, Commerce Data, Protocol Support, Security & Trust, and Technical Performance. The geometric mean produces a single score that penalises gaps — a site cannot compensate for missing product data by having fast page loads.

Inside ChatGPT, the scanner works conversationally. Ask it to scan a site, and it returns the overall score, a breakdown by dimension, and specific findings. Ask follow-up questions — “Why did Structured Data score low?” or “What should I fix first?” — and it walks through the results in context.

It uses the same methodology we built against a corpus of more than 3,500 e-commerce sites across 17 markets, running on a scoring engine certified at v1.0.13 with 177 tests.

Why this matters

The question we keep hearing from CTOs and e-commerce leads is not whether AI agents will start transacting on behalf of consumers. It is when, and what they should be doing about it now.

Most of them do not have a good answer. The major retailers certainly don’t. Shopify’s own storefront scores 45 out of 100. Walmart manages 35. Nike sits at 25. Etsy scores 9. (Benchmark scores as of March 2026; scan any site for its current score.)

These are not obscure sites. They are the largest e-commerce operations on the planet, and none of them are ready for agent-mediated transactions.

And the infrastructure gap is wider than most people realise. Protocols like UCP, ACP, and MCP exist. Payment rails are being built by Visa, Mastercard, and PayPal. But adoption on the merchant side has barely started. Our scans show an average score of 21.2 across the corpus, with a median of 21. The highest score in the dataset is 49 — not even halfway to what full readiness would look like.

Having the scanner inside ChatGPT puts the assessment where the conversations are already happening. Product managers evaluating their competitive position. Developers researching agent commerce protocols. Consultants scoping readiness projects. They can now check a score in the same window where they are doing the rest of their thinking.

What you get

A scan from ChatGPT returns the same data as a scan on zeodyn.com:

• Overall Zeodyn Score (1–100) with readiness band: Not Ready, Limited, Developing, Strong, or Agent-Ready
• Six dimension scores showing where a site is weakest
• Key findings — what the scanner detected and what was missing
• Actionable context — ChatGPT can explain the results, suggest priorities, and answer follow-up questions about any dimension

The ChatGPT scan assesses the homepage. For multi-page scans — which include automatic product page discovery and typically produce higher scores — use the scanner at zeodyn.com/scan with a Pro or Growth subscription.

Not just ChatGPT

The GPT Store listing is one part of a broader push to put the scanner where people and agents already work.

Zeodyn Scanner is also available as an MCP server for Claude Code, so developers can scan sites directly from their terminal. A Chrome extension is in review on the Chrome Web Store — once approved, it will show agent commerce readiness scores as you browse.

On the API side, the full REST endpoint at /api/v1/scan remains available for programmatic access, with documentation at zeodyn.com/docs.

We are building Zeodyn to be consumed by both humans and AI agents. The scanner’s own site scores 87 out of 100 — Strong band — because we are implementing the same standards we assess others against. That includes ACP checkout support, MCP tool exposure, structured product data, and full protocol compliance.

Scan something

The best way to understand what the scanner does is to point it at a site you care about.

Open Zeodyn Scanner on ChatGPT →

Or scan directly at zeodyn.com/scan.