What it is
The classifier walks a user through three plain-language questions drawn from Annex III of Regulation (EU) 2024/1689 (the EU AI Act). Based on the answers, it returns one of three editorial verdicts: Likely High-Risk, Possibly High-Risk, or Likely Not High-Risk. Each verdict links to the relevant Article 26 or Article 27 guidance on agentliability.eu.
The widget is intended for operators, compliance officers, and advisors who want a quick first-pass read on a specific deployment. It is an editorial tool. It does not constitute legal advice.
How to embed
Paste this single snippet wherever you want the classifier to appear:
- Width
- Set to 100% for responsive containers. The widget is usable from 320 px to 800 px wide.
- Height
- 500 px covers the three-question flow and verdict. If your layout is very narrow, 560 px gives more breathing room.
- Dependencies
- None. The widget is fully self-contained. No Google Fonts, no external CSS, no external JS. It loads from agentliability.eu directly.
- Data transmitted
- None. The classifier runs entirely in the browser. No form data is sent anywhere.
- File size
- Under 25 KB. Loads in under 200 ms on a standard connection.
Live preview
The widget below is live. It is the exact embed that will appear on your site.
License
CC BY 4.0
The widget is published under the Creative Commons Attribution 4.0 International licence. You may embed it freely on any site, commercial or non-commercial, without modification or with modification, as long as the attribution built into the widget footer remains intact. We ask only for the brand credit that is already part of the widget.
We do not require a separate backlink from the host page. The "Powered by Agent Liability EU" line in the widget footer is sufficient. If you choose to add a text credit on your page, we welcome it, but it is not a condition of use.
Updates
The classifier follows the official EU AI Act text and Annex III. It will be updated when the European Commission, the AI Office, or the standardisation bodies publish interpretive guidance that materially changes the scope of the high-risk categories. Because the embed loads the file directly from agentliability.eu, your embed will receive updates automatically without any change to your code.
The widget footer includes a version reference in the embed URL. If we ever release a version with structural changes, we will host the prior version at its original URL to ensure backwards compatibility.
Help and feedback
If the classifier returns an unexpected result, if you have found a factual error in relation to Annex III, or if you have a suggestion for improving the question logic, write to the editorial address: editors@agentliability.eu
We will acknowledge substantive feedback within five working days. If a correction is warranted, we will note the change in the version history on this page.