Classify your AI risk
with the reasons recorded
The EU AI Act sorts every AI system into prohibited, high-risk, limited, or minimal, and each tier carries very different obligations. Valdra's guided classifier walks you through Article 5, Annex III, GPAI, and Article 50 to produce a classification from your answers, with the obligations attached.
distinct obligations trigger the moment a system is classified high-risk
1 EU AI Act, Title III, Chapter 2
Try the free Risk CheckerEvery AI system, registered and tiered.
Register every AI system you build or buy and find its EU AI Act risk tier by answering a short questionnaire. Valdra then lists what that tier requires.
Request a demoDocumentation drafted from your registry.
Draft Annex IV technical files, model cards and Fundamental Rights Impact Assessments from your registry data. When a system's model changes, Valdra flags the system for reclassification.
Request a demoAdditional features
Request a demoGuided Classification
A structured questionnaire walks each system through prohibited practices (Art. 5), high-risk use cases (Annex III), GPAI thresholds, and Art. 50 transparency.
Clear Tier, Reasons Recorded
A clear tier (prohibited, high-risk, limited or minimal) with the reasons and statutory basis recorded for your governance file.
Obligations Attached
The moment a system is classified, the obligations that tier triggers are surfaced: conformity assessment, human oversight, documentation, registration.
GPAI Detection
Identifies general-purpose AI and the additional transparency and systemic-risk obligations that apply to foundation models.
Reusable Engine
The same classification engine powers Valdra's public Risk Checker: classify a system once, reuse the result across documents and obligations.
Re-Classification Flags
When the model or version changes, the system is flagged so you can classify it again.
Learn more about Valdra
Close your gaps and build trust
Start free, no credit card.
🇨🇦 Records stored in Canada · Bilingual