Our apps and the EU AI Act: what changes, what doesn't

August 3, 2026 · Privacy · AI Act · European Union

🇪🇺

AI Act · Transparency & Privacy

As of August 2, 2026, part of the EU's Artificial Intelligence Act (AI Act) obligations became applicable. Some of our apps use AI; others don't. Here's what that means, app by app, in plain language.

I. What the regulation says, in one sentence

The AI Act sorts AI uses by risk: prohibited, high-risk (hiring, credit scoring, law enforcement, biometric identification…), limited risk — with a transparency requirement — and minimal risk. None of our apps fall into the "high-risk" category. Wherever AI is present, our only obligation is to clearly inform users they're interacting with one.

II. BookScout & Voyel: when you're talking to an AI, we tell you

BookScout's AI Assistant (RAG over your library, choice of provider — Gemini, Mistral, OpenAI, Anthropic, or Ollama locally) and Voyel's AI document extraction (Apple Intelligence by default, cloud providers optional) both show a warning that must be accepted before first use: which provider receives your data, and that API costs/limits are your responsibility. That's exactly the transparency the regulation asks for — nothing changes on our end, it was already there.

III. HomeCamCheck: detecting presence, not identifying anyone

HomeCamCheck detects human silhouettes via Apple's Vision framework, entirely on-device: no facial recognition, no person identification, nothing sent to a cloud. The regulation targets real-time biometric identification in public spaces or biometric categorization at work; a home security camera that tells you "someone is present" doesn't fall in that scope.

IV. OneLedger: zero AI, zero automated advice

OneLedger's projections, simulations, and Expert Mode all run on deterministic calculations (amortization, taxation, M2/M3 money supply) — no model, no learning involved. OneLedger therefore falls outside the regulation's scope. The app's legal notice already states it provides no investment, tax, or legal advice — a stricter standard than the AI Act requires.

Conclusion

None of our 4 apps offers a high-risk feature under the regulation. Where AI is used, it's with your explicit consent and the transparency the law requires; where it isn't, that stays true after August 2, 2026 just as it was before.