Skip to main content

About Auryth

Swiss-incorporated · EU-hosted · Built by domain professionals

Our story

Auryth was born from frustration — not from a pitch deck.

Our founding team spent years working in regulated professions — tax advisory, accounting, enterprise consulting — where getting an answer wrong has real consequences. We know what it's like to spend an hour navigating dense legislation, cross-referencing circulars, and chasing down rulings across fragmented government databases, only to be 90% sure of a single answer.

We're also deeply technical. When we tested general-purpose AI on real professional questions, we saw what any experienced practitioner would see: confident answers that were subtly, dangerously wrong.

Outdated rules. Blended jurisdictions. Missing exceptions. No sources.

So we built something better. Not a chatbot that guesses — a research platform that thinks the way a professional thinks. Methodically. With sources. With awareness of what changed and when.

"In regulated domains, 'probably right' isn't good enough."

We started with Belgian tax law — one of the most complex fiscal systems in Europe — and proved the approach works. Now we're expanding to other regulated domains where the same problems exist: dense regulations, temporal versioning, contradictory sources, and zero tolerance for error.

What we see in the market

An honest observation — not a competitor attack.

Most tools today

Enhanced search engines

Accept natural language input, run a search, return documents — sometimes with an AI summary on top. Useful, but not research.

What research actually means

Structured reasoning with sources

Understanding which sources outrank others. Knowing when a rule was superseded. Recognizing contradictions — and saying so.

When the stakes are professional liability and client trust, "probably relevant" isn't the same as "authoritative." That's the gap we're building to close.

Our approach

Most AI tools give confident answers. We think that's dangerous.

Honest by design

Every answer comes with a confidence score. When we're sure, we show why. When we're not, we say so.

Full transparency

See which sources informed an answer, how they were ranked, and why certain provisions were set aside.

Your judgment first

We'd rather give you an honest "we're not sure" than a confident answer built on thin evidence.

Where we work

Each vertical is built from scratch with domain-specific data structures, retrieval logic, and validation. We don't bolt plugins onto a generic model.

Auryth TX AI

Pre-launch

AI-powered Belgian tax research. Reads legislation structurally, tracks temporal versioning, and scores every answer with confidence levels and source citations.

Visit auryth.ai/en/tax/ →

Auryth VetPharma AI

Coming 2026

AI-powered veterinary pharmacovigilance. MRL verification, withdrawal period tracking, species-specific safety data, and regulatory currency.

Company

Auryth AG

Swiss corporation · Canton of Schwyz

We chose Switzerland for its strong data protection tradition, political stability, and bidirectional data protection adequacy with the EU. All infrastructure is hosted in the EU.

EU-hosted (AWS Frankfurt)
GDPR-compliant
SOC 2 Type II on roadmap
No model training on your data

Interested in investing?

We're building the research layer for regulated professions. If you're an investor in vertical SaaS, legal tech, or AI infrastructure, we'd welcome a conversation.

Learn more

Ready to try it?

Our first product, Auryth TX AI, is in pre-launch. Join the waitlist for early access to AI-powered Belgian tax research.

Join the waitlist