How scans and the purchase-journey test work
Each scan launches an isolated Chromium browser (1280×800, English locale, reduced motion) that identifies itself with the TabToCartBot user agent.
Journey steps
| Step | What happens |
|---|---|
| Homepage | Load, run automated rules, walk the page with the Tab key |
| Browse | Open a category/listing page found in the navigation or sitemap |
| Product | Open a product page, detect the add-to-cart control |
| Add to cart | Select default options if required, focus the add-to-cart button and press Enter. If Enter does nothing but a click works, that is reported as a keyboard failure. Cart drawers are checked for dialog role, name, focus and Escape |
| Cart | Open the cart page and detect the checkout control |
| Checkout | Activate the checkout control with Enter and test the information step |
| Form errors | Submit the information step empty, or type an invalid email and move focus away, then check how errors are exposed |
Safety rules
- No payment data is ever entered.
- Buttons whose text matches pay / place order / confirm order (in 10+ languages) are never activated.
- No account is created and no personal data is typed.
- If a bot challenge or login wall appears, the step is reported as blocked and the test stops there.
Automated rules
On each page we run axe-core with the WCAG 2.0/2.1/2.2 A and AA rule sets plus selected best practices, and our own checks: keyboard focus visibility, keyboard traps, focus on invisible elements, add-to-cart and checkout operability, cart drawer behaviour, autocomplete on personal-data fields, required-field semantics and error announcements.
Scan types
- Baseline — first scan after adding a store.
- Daily journey — the 7 journey steps only.
- Weekly full — journey plus additional pages up to your plan limit (search, account, more categories and products, key content pages, and any URLs you add).
- Manual — on demand, limited per day.