Terms

These terms cover this site and its API. The models have their own licences, listed below.

Research use

OpenMorals is a research tool. A score is a model's reading of a piece of text. It is not advice, not a diagnosis and not a finding about any person. Read the limitations, including the uses to avoid.

No warranty

The site, the API and the models are provided as they are, without a promise of accuracy, availability or fitness for a particular purpose. The API can change, slow down or stop.

Using the API

• A key is for the app it was issued to. Keep it secret, and don't share it.
• Stay within the limits in the API guide. A key that is misused can be revoked.
• Don't send text you have no right to send, and don't use the scores for the purposes listed under uses to avoid.
• Don't disrupt the service or try to get access you haven't been given.

Licences

• OpenMorals RoBERTa: MIT
• OpenMorals Gemma: Apache-2.0 for the adapter. The base model has its own terms.
• Training data: Moral Foundations Reddit Corpus (MFRC), under CC BY 4.0. It stays under its own licence.

A licence to the weights doesn't grant the right to use the OpenMorals™ or AI Attributes™ names, or imply endorsement.

Changes

These terms describe the site and the API as they are now, and may be revised when they change.