HabichtLLM (sovereign)
Open-source/open-weight models hosted sovereignly in Austria/the EU – first choice for sensitive content. Custom models are possible on request, including training.
HabichtLLM is the sovereign model and compute layer behind Habicht: operated in Austria and the EU, multi-model, and built so that your data does not flow out to external models. Usable as part of the Habicht platform – or as pure, sovereign AI compute for your own applications.

Habicht is not tied to a single provider. Depending on confidentiality and task, the platform picks the right model.
Open-source/open-weight models hosted sovereignly in Austria/the EU – first choice for sensitive content. Custom models are possible on request, including training.
Connection to EU models such as Mistral for more performance – within the European legal area.
Available for tasks involving generally accessible knowledge anyway – on a contractual basis (data-processing agreement) and only with the protection mechanisms below.
So that nothing sensitive leaves the house when external models are added, several mechanisms apply – you keep control per use case.
Sensitive content goes exclusively to sovereign or local models.
Protection mechanisms prevent sensitive data from being sent to external models.
For each task it is defined which model may see which content.
HabichtLLM runs on infrastructure in Austria and the EU – no default to the US cloud. This lets you meet data-sovereignty and compliance requirements at the compute level already.
You only need sovereign AI compute – as model access for your own applications? HabichtLLM can be used standalone. Talk to us about sovereign inference capacity.
Within Habicht, HabichtLLM sits behind every workflow and every answer.
Whether as part of the platform or as pure inference capacity: we'll find the right sovereign solution.