Newtresor: one sign-in, every DuckDB secret your role allows

DuckDB, ready for your organisation.

Hugr Lab builds open DuckDB extensions for the systems companies already run, and the pieces that make DuckDB safe to use across a company. Built in Germany, in the open.

Hugr Lab
DuckDB extensions · data engineering
WHY NOW

DuckDB is becoming company infrastructure.

In August 2026 AWS agreed to acquire DuckLabs, the team behind DuckDB. DuckDB itself stays open source. Companies in Europe that adopt it still need it to fit their systems, their identity providers and their auditors, from a vendor under EU jurisdiction. That is what Hugr Lab works on.

PRODUCTS

Products for DuckDB in production.

Each product works on its own and installs in minutes. Open source at the core; run it yourself or let us run it in your cloud.

Open source + BUSLNew
tresor

Secrets for DuckDB

One sign-in through your identity provider, and every analyst and application gets exactly the credentials their role allows. Rotate and revoke centrally, or keep secrets in the vault you already have.

About tresor
Open sourceAvailable
mssql and mssql_ducklake

SQL Server for DuckDB

Query and load SQL Server, Azure SQL and Fabric from DuckDB over the native protocol. Keep a DuckLake catalog in SQL Server.

About the SQL Server extensions
Open sourceAvailable
hugr classic

GraphQL over your data

The GraphQL engine you may know from hugr-lab.github.io. One API over databases, files and services.

About hugr classic
HOW WE BUILD

Built for organisations whose data cannot leave.

Runs where your data is

Everything we ship runs on your machines or in your own cloud tenant. Nothing is copied to us.

Readable code

Our extensions are open source (MIT, Apache-2.0); the few commercial parts are source-available. You can read everything you run.

European jurisdiction

Hugr Lab is based in Germany. Azure first, with Open Telekom Cloud and STACKIT next.

MEASURED IN PUBLIC

Built in the open, measured in public.

We publish how our extensions work and how fast they are. Start with SQL Server: two lines in DuckDB.

DuckDB
INSTALL mssql FROM community;
LOAD mssql;
ARTICLEFrom 51 seconds to 1 second: DuckDB + SQL Server, 49x fasterARTICLESame hardware, two catalogs: DuckLake on SQL Server and PostgreSQL, measured in AzureARTICLEWhat SQL Server wants from a lakehouse catalog: seven measurementsARTICLEQuerying Azure SQL and Microsoft Fabric from DuckDB with Entra ID

Need a hand with your data platform?

Data engineering in short projects: SQL Server into a lakehouse, DuckDB in front of your BI, or a platform that stays in your cloud. Done by the people who wrote the code.

Hugr Lab UG is being founded. We take on projects from November 2026.

Write to usAbout Hugr Lab