Where this comes from
Article 50(4) and Article 3(60) of Regulation (EU) 2024/1689, the Commission guidelines on transparency of AI-generated content, the Code of Practice (Section 2, deployers), and the Commission's Article 50 FAQ.
Not official
Independent and unaffiliated. Not endorsed by the European Commission or the AI Office, and not a signatory to the Code of Practice — using this tool doesn't signal adherence to it.
Private by design
No account, no cookies, no third-party requests. Your answers never leave your browser. Visits are counted with self-hosted, cookieless analytics that cannot identify you, and you can switch that off on the privacy page.
Accessibility
Built to WCAG 2.1 level AA, the standard EN 301 549 and the European Accessibility Act point to. Full keyboard operation, visible focus, text alternatives throughout, contrast checked against AA thresholds, and no information carried by colour alone. Found a barrier? Please open an issue — it will be treated as a bug, not a nice-to-have.
Open source
MIT licensed. The decision logic is one readable file — corrections and pull requests welcome.
Take it with you
Prefer to run the check inside an AI assistant you already use? Same questions, same conservative bias against over-labelling, same official sources — packaged two ways. It's a snapshot, so the tool above stays the authoritative version.
Download options and where each one works
File
Works with
How to use it
label-or-not.skill
Claude only
Install as a Claude Skill — triggers automatically when it's relevant
label-or-not-instructions.md
Claude, Copilot, Gemini, or any chat-based assistant
Paste into custom instructions, a project/Gem, or straight into chat
Free, and generated with support of AI tools
This tool is free, has no ads and collects nothing. It has been generated with AI and supporting tools. If it saved you some time, you can buy me a coffee .
Imprint
Privacy
Deutsch