About
Euraika-Labs is where Euraika writes about what it builds.
The systems — Aegis, Hermes, Vane, and the Pan family — are sovereign-by-default AI infrastructure for organisations whose data, compliance posture, or jurisdiction makes US-cloud LLM stacks a non-starter.
They ship through Euraika's product lines. The open source lives at github.com/Euraika-Labs. The writing about how we built them, what we got wrong, and what we'd do differently lives here.
We write in a single institutional voice. Posts are not personally bylined. Credibility comes from what we ship and what we measure, not from titles.
What we work on
Aegis
AI-powered compliance command center.
Single-tenant SaaS for security auditing, evidence management, and continuous compliance under SOC 2, ISO 27001, NIS2, and the EU AI Act. Each customer gets their own Kubernetes namespace.
Hermes
Production LLM gateway.
The disciplined hand-off between an application and a model provider — multi-runtime, with a model catalogue, governance hooks into Aegis, and a routing layer that respects EU jurisdiction by default.
Vane (Camiel)
Privacy-focused AI answering engine.
A search and answer engine with verified synthesis and citations. Runs on top of SearXNG; never sends a user's query to a third-party search index.
Pan
Self-hosted AI workspace.
Open source. Desktop agent with full PC control, web UI, native shell, and a 3D engine — all on the operator's hardware. See the project page.
How we're set up
On-prem Kubernetes in EU jurisdiction. GitOps via ArgoCD from self-hosted GitLab. Errors flow to a self-hosted Sentry. Analytics is self-hosted Matomo, honouring Do Not Track. The site you are reading runs on the same cluster as the products it talks about.
The people
Behind the institutional voice is a small team of engineers and researchers who build this stack because we love the work — and because we think Europe deserves AI infrastructure that answers to Europe.
- Bert Colemont
- Andrei Lavrenov
- Laurence Bedoret
- Corey Stables
- Amin Zakkane
We research together, ship together, and argue about the details — for a better AI future for Europe.
Common questions
- What is Euraika-Labs?
- Euraika-Labs is the editorial surface for Euraika, a European AI research lab. The lab builds sovereign-by-default AI systems — Aegis, Hermes, Vane, and the Pan family of open-source projects — for organisations whose data, compliance posture, or jurisdiction makes US-cloud LLM stacks unworkable. This site publishes the writing about how those systems are built.
- Where is data hosted?
- On-prem Kubernetes in EU jurisdiction. Sovereignty under our definition requires that the legal, contractual, and operational stack all sit inside the same regime — not just data residency on hardware operated by a US-headquartered provider.
- Are the products open source?
- The open-source side lives at github.com/Euraika-Labs. Pan, Hermes-Agent, ai-slopcheck, and omi-regolo-integration are open. The commercial products — Aegis (compliance) and Hermes (LLM gateway) — ship under commercial licences, but are designed to run end-to-end inside the customer's own infrastructure.
- Who runs the lab?
- A small team of engineers and researchers: Bert Colemont, Andrei Lavrenov, Laurence Bedoret, Corey Stables, and Amin Zakkane. We post in a single institutional voice — credibility accrues to the lab, not to individual bylines.
- How do I get in touch?
- hello@euraika.net.