Free · No signup · 30 sec · deterministic

Will your AI agent carry the evidentiary logging signals the EU AI Act requires when high-risk logging obligations bite by December 2027?

Article 50 AI-content transparency obligations apply from August 2, 2026; the high-risk Article 11/12 logging obligations phase in by Dec 2, 2027 (Annex III) and Aug 2, 2028 (Annex I) under the Digital Omnibus timeline (provisional until published in the Official Journal). Penalties reach €35M / 7% of global turnover. Article 12 requires automatic, tamper-evident event logs over the lifetime of every high-risk AI system — mutable JSON traces don't satisfy "automatic recording." Insurers now gate model-performance warranties (Armilla, backed by Swiss Re / Greenlight Re / Chaucer; Lloyd's coverholder) on non-repudiable proof of behaviour. Paste your agent endpoint — we'll deterministically run 17 evidentiary checks across 6 EU AI Act articles (Art. 11/12/13/14/26/43/72) and show you what the same trace looks like as a Catalyst-Brain .rain packet (cryptographically signed, sub-100KB, replayable).

Anonymous. We fetch the URL once, run 17 evidentiary checks deterministically (regex / JSON-path — no LLM), return a shareable report.
Scope. This audit runs 17 evidentiary-signal checks across 6 EU AI Act articles (Art. 12 logging plus the Art. 11/13/14/43/72 substrate obligations) — roughly 6 of ~14 provider obligations for high-risk AI systems (Annex III). A 17/17 score means the canonical signal patterns are present; it does not mean the system is "EU AI Act compliant" overall. Full compliance also requires quality management, fundamental-rights impact assessment, conformity assessment, post-market monitoring, and EU registration — work your lawyers + compliance team own.