Your Data, Your Infrastructure

A container with 20 EU sources, pre-annotated and ready to query. Deploy anywhere.

20 EU sources 28 national databases 2.4M terms 24 languages 290K cross-references 21 topic domains

What You Get

Full ownership A container you deploy on your servers, your Azure tenant, or your private cloud. Runs independently, stays under your control, works on your schedule.
All sources, pre-annotated 20 EU institutions. 28 national law databases. 290,000 legislative cross-references. 2.4 million multilingual terms. Pre-tagged with topic classifications, legal modality, and procurement codes.
Complete privacy Your queries stay on your infrastructure. The connector runs offline, stores everything locally, and operates independently. Your data, your sovereignty.

How It Works

1 Pull the container docker pull pauhu/data:latest
2 Configure your sources Select topics, languages, and update frequency
3 Query REST API on port 8003, or connect via Eclipse Dataspace

Why Sovereign

No per-seat AI tax

Enterprise AI add-ons charge per seat, per month, forever. For a 500-person agency, that is hundreds of thousands per year — for a tool most employees reject. Industry data shows 96.7% rejection rates across 450 million seats.

Pauhu is a one-time container deployment. Works with your existing tools. No vendor lock-in, no recurring per-seat fee. Pay for what you use, not for what you deploy.

Supply chain sovereignty

Foreign governments have designated AI providers as supply chain risks and revoked access with a single policy decision. If your regulatory intelligence depends on a service outside EU jurisdiction, it can be switched off without notice.

Pauhu runs in your browser, on your server, under EU jurisdiction. No foreign government can shut it off. Built for NIS2 agencies, ministries, and EU institutions that need continuity guarantees.

Runs on the hardware you already have

Cloud AI inference requires expensive GPU clusters. DRAM prices have risen 23% in a year. Every query you send to a cloud model costs money and creates a dependency.

Pauhu runs on devices with as little as 4 GB RAM. The generation model is 300 MB — it fits on an 8 GB laptop. No cloud required. No GPU required. Your data never leaves your device.

Requirements

Runtime: Docker, Podman, or any OCI-compatible runtime

Memory: 4 GB minimum, 8 GB recommended

Storage: 50 GB for the full annotated dataset

Connected container: Azure Arc for metered on-site deployment

Deploy the connector See pricing

30 days free · No credit card required

Eclipse Dataspace Connector · LDS Protocol · Azure Marketplace (MACC-eligible)