AccessureBot

It opens a page the way a browser would, waits for it to render, and runs three open-source accessibility engines over the result: axe-core, pa11y and IBM Equal Access. It reads what any visitor can read. It does not log in, does not submit forms, does not touch anything behind authentication, and never requests private or internal addresses.

A scan is a handful of page loads, images and fonts skipped. Public demonstration scans are limited to three per address per day. If it ever costs you more than that, tell us and we will stop.

To measure a page, the engines have to run inside it, and to inject them the scanner switches off that page’s Content Security Policy— the rule a site sets about which outside code its own visitors’ browsers may run. This happens only inside our own throwaway browser, on our own machine, for the seconds the measurement takes. Nothing is changed on your server, no protection is weakened for anyone visiting your site, and no data of yours is touched. Every accessibility testing tool works this way; we would rather you read it here than discover it in a log.

For a customer’s own sites: the findings, in their account, private (Terms 8b). For a demonstration scan of a public home page: a summary — a score, counts by severity and a few example findings — cached at accessure.app/scan/<domain>. That page is excluded from search-engine indexing, never sold and never shared.

Email support@accessure.app with your domain. We remove the cached page and add the domain to a permanent do-not-scan list; no reason needed and no reply required from you beyond that one message.

You can also block it yourself in robots.txt:

User-agent: AccessureBot
Disallow: /

Accessure, operated by Matteo Salmaso, Italy. Questions: support@accessure.app.