Sovereign Managed Inference

We are currently in a controlled rollout phase. To ensure consistently stable latencies and sufficient capacity, we grant access selectively to selected partners first. After registration via the form, we review your usage profile and enable access step by step. Production-ready workloads are the stated goal of our rollout, not the exception.

Yes. We operate exclusively our own data centres in Europe and do not use any rented or outsourced third-party capacity. Your data never leaves the EU, and there is no dependency on external cloud providers.

We provide a continuously growing selection of open-source models through a unified API. Deploying your own fine-tunes or custom checkpoints is planned for a later expansion stage and is currently in development. We are happy to inform interested partners as soon as initial capacity is available for this.

Capacity is not allocated manually but provided via automated autoscaling on our own infrastructure. Since we own the entire stack — from rack to API — we can scale capacity immediately, regardless of a third-party provider’s availability or pricing.

These components — RAG pipelines, MCP integration, as well as function-calling and agent support — are in active development and will be gradually enabled for partners. If this is relevant to your use case, we recommend noting it in the "Expected Usage" field of the form to support prioritisation.