SORI

Training Data Use Agreement

AI-MIDI Web Transcription Training Data Use Agreement

Effective
2026-07-28
Displayed language
English · original

The English document is the original. Localized translations are provided for convenience.

Required Agreement

This agreement is required to use Web Transcription. If the User does not agree, Web Transcription is unavailable. This agreement does not apply to the AI-MIDI API.

Purpose and Scope

The User agrees that the Company may anonymize Web Transcription Inputs and Outputs and use them to train, evaluate, validate, measure, and improve AI-MIDI models.

Processing

For these purposes, the Company may store, copy, transform, separate, analyze, and incorporate Inputs and Outputs into training datasets.

Segregated storage

Training copies are managed separately from operational service data and processed so they are not linked to account-identifying information.

After Account Withdrawal

Operational Inputs and Outputs are deleted on withdrawal. Copies anonymized and incorporated into a training dataset before withdrawal may continue to be used for training, evaluation, and improvement of new models. The Company does not guarantee deletion of the data's influence from an already trained model, model rollback, or retraining.

User Confirmation and Responsibility

If an Input contains personal information, the User confirms that the User has a lawful basis and all rights and permissions necessary for the Company to process that information for the service and training purposes described in this Agreement. The User must not submit personal information that the User is not authorized to provide. The User bears all disputes and responsibility concerning copyright, neighboring rights, personal information, privacy, and other third-party rights in connection with Inputs and Outputs.

Contact

For inquiries about training-data use, contact team@sori-ai.com.

team@sori-ai.com