Does your chat widget disclose that it's AI?
Paste a URL or page source. You get observations about what's there, with the paragraph of the Commission's guidelines behind each one. No sign-up to run the scan.
Observations
What this scan cannot tell you
The four checks that actually matter
Two minutes, in your own browser. These test the thing a scanner cannot see.
- Open the chat in a private window. Is the disclosure visible before you type anything? Article 50(5)
- Read it. Does it contain "AI" or "artificial intelligence"? "Assistant", "powered by…" and "smart replies" do not count. Guidelines para (38)
- Tab to it. Is it reachable and announced by a screen
reader? Is it
aria-hidden? Article 50(5) - Ask the bot "are you a human?" Does it say it is an AI? Guidelines para (40)
Check 4 fails on almost every default installation.
Want the fix for your platform?
The install recipe for the widget we detected, the disclosure text in English and Polish, the system-prompt snippet that makes check 4 pass, and a note when the guidance changes. Free.
This is not a compliance check, and it never says "pass".
A scanner reads the HTML a server returns. Chat windows are drawn by
JavaScript after the page loads, often inside an iframe — so no scanner can
see inside the chat window, which is exactly where Article 50(1) requires the
disclosure to be. Absence of any AI wording is a strong signal. Presence of it
proves nothing. Nobody can tell you whether you are compliant from a URL, and
anyone who says otherwise is selling something.