Build AI capability you can keep control of.
We are preparing a service to help you find, evaluate and use AI around your goals, with clear choices about your data, your models and where they run.
In partnership with BuiltAPI · DataClam →
Coming soon. The model archive, evaluations and weekly matching service are planned. Register for updates and a conversation about your needs. Launch timing will be shared when confirmed.
Register your interest.
Receive launch updates or tell us how you would like to contribute. Your registration helps us understand which applications to prioritise.
WINNIIO × BuiltAPI
We are developing this initiative in partnership with BuiltAPI, the team behind DataClam.
We are starting with our own need for continuity: preserving access to selected AI model versions and the knowledge needed to use them. Our aim is to bring that capability to customers through a curated archive, model selection, evaluation and deployment advice.
BuiltAPI has proposed hosting the archive on its storage in Germany and managing the model catalogue in DataClam. WINNIIO leads the advisory initiative, connecting model capabilities to the outcomes people and organisations want to achieve.
Keep your options open as AI changes.
Our starting point is the person or organisation using AI: their goals, their context and their authority to decide. We want useful knowledge to build over time, with access granted on their terms.
That ambition connects personal research, manufacturing and physical assets. Each needs its own evidence and safeguards. The shared question is who controls the data, the tools and the next decision.
Preparing now means defining what you need to preserve, which models you are entitled to use and how a capability could move between environments. A documented archive and reproducible evaluations are parts of that preparation.
Sovereignty through practical control.
Our proposed assessment would document the controls an application needs and how to test them.
- Data and access
- Decide what is collected, where it is processed, who can access it and how permission is withdrawn.
- Models and permissions
- Record exact versions, provenance and licence terms, including what can be retained, adapted and redistributed.
- Deployment and continuity
- Assess local or controlled cloud operation, provider dependencies, export, recovery and the ability to change suppliers.
- Evidence and accountability
- Evaluate the intended task, document uncertainty and costs, and keep human authority over consequential actions.
Keeping a copy of model weights addresses availability. Sovereignty also requires enforceable permissions and operational controls across the application. Open-weight models have individual licence conditions; application suitability needs separate evaluation.
Further reading: European Commission: technological sovereignty · Open Source Initiative: the Open Source AI Definition
What we are preparing.
- A curated model archive
- Selected, documented model versions with measured storage needs and licence review.
- Selection and evaluation
- A shortlist matched to your decision, followed by a scoped comparison with your current approach.
- Weekly project matching
- A review of relevant releases and changes in your project, with reasons to investigate or keep your current approach.
- Community contributions
- Model suggestions, reproducible tests and practical experience, with attribution and review.
The planned commercial model separates advisory and implementation fees from agreed infrastructure costs. Any paid work, storage or compute would require a separate scope and budget.