Free tool
Agent Readiness Scanner
Can an AI agent find your plans, parse your pricing, and reach checkout? Enter your URL and find out in seconds — a real crawl of your site, graded like a report card.
Why agent readiness matters now
AI assistants have started completing purchases, not just recommending them — OpenAI and Stripe published the Agentic Commerce Protocol, and agent-initiated checkout is live in ChatGPT. Before any of that machinery can transact with you, an agent has to do something much more basic: fetch your site, find your plans, and understand your prices. That's the layer this scanner tests.
Most subscription sites fail it today — prices rendered only by JavaScript, checkout behind un-followable buttons, cancellation policies an agent can't locate, and robots.txt files that turn away purchasing agents along with training crawlers. Every one of those is fixable, and the full report tells you how.
What the scan checks
Pricing page discoverability
Is there a crawlable link from your homepage to your plans, and does the pricing URL answer?
Machine-readable offer data
Do your plans carry schema.org Product/Offer JSON-LD with real prices, or do prices exist only as display text — or only after JavaScript runs?
Checkout reachability
Can an agent follow a plain link from plan to checkout, or is the only way in a JavaScript button it can't click?
Cancellation path
Is there a discoverable link explaining how subscribers cancel — an agent-trust signal and, increasingly, a compliance one?
Upgrade / plan-change path
Can an agent find how existing subscribers move between plans?
AI-agent access (robots.txt)
Does your robots.txt block the user-agents of ChatGPT, Claude, Perplexity, and other assistants site-wide?
llms.txt guidance file
Do you publish an llms.txt pointing language models at your pricing, docs, and policies?
How it works
- 1
Enter your site's URL
Homepage or pricing page — the scanner normalizes it and crawls from there.
- 2
We crawl what an agent sees
The scanner fetches your homepage, robots.txt, llms.txt, and pricing page server-side and reads the raw HTML — the same first impression an AI agent gets.
- 3
Read your grade
Seven deterministic checks roll up into a letter grade and a 0–100 score. Every check shows pass, warning, or fail.
- 4
Unlock the full report
Enter your work email to see the evidence behind each finding and a concrete fix for every failure.
Frequently asked questions
- What is agent readiness?
- Whether an AI agent — ChatGPT, Claude, Perplexity, or a shopping assistant acting for a customer — can actually do business with your site: find your plans, parse your pricing, reach checkout, and locate cancel and upgrade paths. Most subscription sites were built for human eyes and JavaScript-heavy browsers, and fail several of these checks today.
- What does the scanner actually check?
- It fetches your homepage, robots.txt, llms.txt, and your pricing page, then runs seven deterministic checks against the real HTML: pricing page discoverability, machine-readable offer data (schema.org JSON-LD), checkout reachability, cancellation path, upgrade path, robots.txt treatment of AI-agent user-agents, and llms.txt presence. Every finding quotes what the crawl found — nothing is estimated or invented.
- Why does the scan only see my server-rendered HTML?
- Because that's what most agents see first. When an AI assistant fetches your page, it typically reads the HTML your server returns — if your prices, links, and checkout buttons only appear after client-side JavaScript runs, many agents never see them. That's not a scanner limitation; it's the finding.
- What is the Agentic Commerce Protocol (ACP)?
- An open specification, published by OpenAI with Stripe, that lets AI agents complete purchases programmatically — it powers agent-initiated checkout in ChatGPT. Being crawlable and machine-readable is the prerequisite layer: an agent has to find and understand your offer before any protocol can transact it.
- My site got a low grade — is it broken?
- Not for humans. A low grade means an agent reading your server-rendered HTML can't find or parse what it needs. The full report lists each failed check with what we found and a concrete fix — typically structured data on the pricing page, plain links to checkout, and a published cancellation path.
- Do you store my site's data or my scan results?
- No. The scan runs on demand, the report is generated for your browser, and nothing about your site is stored server-side. Your email is only requested if you want the full report, and goes to our CRM like any other form on this site.