CONTRIBUTION PRIVACY PLAN · 22 SEPTEMBER 2026

Our model.
Your choice to contribute.

Optional examples from customers running Viking Rise Automation will help Young Studio build its own VRA model. You will not be asked to train a model or manually assemble a dataset. This product privacy plan is separate from the waitlist and email privacy notice, licensing and normal farm operation.

What you would be helping build

We are building Young Studio’s own VRA model. We will train and evaluate it using teacher-labelled development examples and optional contributions from customers running VRA. Customers contribute examples—not train or maintain separate models themselves.

The experimental VRA vision/language model is planned for a later iteration—not the initial launch. We will release it only after internal testing demonstrates an improvement over the script-based automation baseline. Training and evaluation will happen offline, and qualified model versions will be delivered through VRA updates. A contributed image will not immediately change your farm’s behaviour.

How contribution will work at launch

At launch, optional dataset contribution is part of the release plan: with explicit opt-in, VRA will select useful examples during normal automation, apply privacy controls and contribute approved examples to Young Studio’s VRA dataset. Manual screenshot preparation will not be required for everyday contribution.

The release flow is: choose your contribution settings; let VRA select eligible examples while your configured workflows run; apply supported local privacy filtering; review or remove queued examples; contribute permitted examples for training, evaluation and failure analysis. Supported redaction, minimised metadata, duplicate filtering and retention controls are part of the release requirements.

Privacy filtering must not be described as guaranteed anonymisation. Any limitations and remaining identifying details will be explained before opt-in. The contribution settings will not grant access to game passwords, licence keys or session credentials.

Separate, default-off permission

Contribution will be off unless you explicitly opt in. Failed or low-confidence examples, successful examples and permission for Young Studio personnel to review contributed images will have separate scopes. Consent will record its version, timestamp and scope, and remain accessible from Settings. Declining will not reduce core automation or affect your licence.

VRA still needs to observe screens to perform your chosen workflows and retain appropriate operational diagnostics. That necessary screen observation is not permission to add those screens to a training dataset. Support sharing and marketing consent remain separate too.

What an example may contain

Screenshots can include character and tribe names, chat, resource amounts, city state and emulator labels. A useful training example may pair the image with the observed state, attempted action, verified outcome and limited timing or workflow metadata. Uncertain results must not be labelled as verified success.

Human review, storage and providers

The purpose is improving Young Studio’s VRA model through training, evaluation and failure analysis—not advertising or training a separate model for each customer. Human access will follow the consented review scope.

Before launch, the final notice must identify where contributed data is stored, any hosting or processing services involved, applicable transfers, retention periods and how deletion works. Those operational details are not finalised on this pre-release page; they must be published before collection opens. Website and email providers listed in the waitlist notice do not receive permission to use desktop images merely because you joined the mailing list.

Withdrawal and deletion

You will be able to withdraw from Settings as easily as opting in. Withdrawal must stop new optional collection and sending, and clear unsent examples. Individual and bulk queue deletion, deletion requests for stored contributions and exclusion from future training are release requirements. Network-dependent deletion must show its actual status rather than claim completion while offline.

These controls concern optional contributions, not your licence, farm configuration or operational run history.

Examples already used in a model

Deleting a source image is not the same as removing its influence from an already-trained model. The final notice must explain how stored copies, future training datasets and affected model versions are handled. We will not promise automatic unlearning or hide that distinction. Customer contribution will not open before the deletion and trained-data handling process is documented and tested.

Questions before opting in

Contact [email protected] about this plan, access or deletion. The final notice requires privacy and legal review before launch; this page is not an active request for screenshot consent. You may also raise privacy concerns with the UK Information Commissioner’s Office.