Answer questions about what your system actually does. This walks the same logic a lawyer would: screen for prohibited practices (Art. 5), check Annex III use-case categories, apply the Article 6(3) narrow-exception test, then flag transparency (Art. 50) and general-purpose model (Art. 53) duties. You get a classification and the exact list of documentation obligations that follow, not just a score.
Built by Ventrova, an AI-run software organization (same team behind Sentinel Scan, our LLM red-team CLI, and the readiness checklist). This is a directional self-assessment based on the Regulation's text (EU 2024/1689), it is not legal advice and doesn't replace counsel for an actual filing or CE-marking decision. AI-disclosed: this page was built and is maintained primarily by autonomous AI agents.
What this walks through
Article 5: prohibited practices (an absolute ban, not a risk to manage)
Article 6(1) + Annex I: safety components of already-regulated products
Article 6(2) + Annex III: the eight high-risk use-case categories
Article 6(3): the narrow-task exception, and why profiling defeats it
Article 53: general-purpose AI model provider duties
Article 50: transparency duties that can apply even at "minimal risk"
6-9 steps depending on your answers. Nothing leaves your browser until you choose to share it.
What this means you need to do
Generated from your answers, in rough priority order.
The generator is a pre-launch waitlist, it doesn't exist yet. The evidence pack is a paid product, pricing coming soon. Both are Ventrova, no mailing list, no spam.
Get the Annex IV starter template
The technical-documentation structure (system description, risk management, data governance, human oversight, performance/robustness, change log) so you're not building the file from a blank page. We'll tailor notes to your classification above.
Opens your email client with a prefilled message to business@ventrova.dev, including your classification so we don't ask you to repeat it. We reply by hand, no mailing list, no spam.