Made in Austria.
Open to everyone.
Built in
Austria
Under the rules of
The European Union
A small European team, working in public
No offshore development shop, no anonymous maintainer account.
Questions about a rollout, a connector you need, or where the project is going next? Write to us — the same address reaches the people who wrote the code.
[email protected]Being European
Plenty of tools claim EU hosting and then ask you to ship your most sensitive data to them anyway. Classifyre is built the other way round: the software travels to the data.
Your data never has to leave the building
Classifyre is software you install, not a service you upload to. The desktop build keeps everything on one machine; the Helm chart keeps everything inside your cluster. There is no vendor tenancy to pick a region for, because there is no vendor tenancy.
No dependency on a non-EU cloud
Nothing in the stack requires a hyperscaler account. PostgreSQL and Kubernetes are the only hard dependencies, so the whole platform runs on an EU provider, on-premises hardware, or an air-gapped rack — the same build either way.
You choose the model, and where it runs
AI providers are configuration, not a hardwired vendor. Point the autopilot and the LLM detectors at a European endpoint, at a model you host yourself, or at nothing at all — the rule-based and local-model detectors work without any provider configured.
We work under the regulations you do
GDPR, NIS2, DORA, and the AI Act are not an export checklist for us — they are the law where we sit. That shapes the product's defaults: audit trails on cases, encrypted credentials, workspace isolation, and telemetry you can switch off.
We are not going to tell you a piece of software makes you GDPR-compliant — no software does that. What we can say is that Classifyre never needs a copy of your data, and that every design decision above was made so a European compliance review is a short conversation instead of a project.
Deployment docsApache-2.0
Permissive, patent-granting, and boring on purpose. Fork it, ship it, run it commercially.
The whole engine
Connectors, detectors, investigations, autopilot, desktop app, and Helm chart — not a stripped demo.
In the open
Issues, pull requests, and releases all happen on the public repository, in public.
Built here. Run anywhere.
Download the desktop app and point it at something real. Nothing is uploaded, nothing phones a foreign cloud, and everything you build carries over when you move to Kubernetes.



