Acceptable Use Policy

This Acceptable Use Policy (“AUP”) applies to all use of the Service and is part of the Terms of Service. It also flows down restrictions required by the providers of the AI models we may use.

1. Prohibited uses

• Unlawful, fraudulent, or infringing activity, or violating others’ privacy.
• Uploading content you lack the rights to, or that is defamatory, harassing, or hateful.
• Attempting to generate advice for emergency, medical, legal, or financial reliance and presenting AI output as professional advice.
• Probing, scanning, or breaching security; circumventing rate limits or access controls.
• Using the Service to build a competing model or to scrape/extract other users’ data.
• Generating content that sexualizes minors, promotes violence, or facilitates illegal weapons/CSAM/terrorism.

2. Model-provider restrictions (flow-down)

Some AI capabilities are governed by upstream model licenses and acceptable-use policies. To the extent any request is served by a model licensed under the Meta Llama 3 Community License and Acceptable Use Policy, you agree not to use outputs in the ways that license prohibits (including the uses listed above), and the same restrictions flow down to your guests. This flow-down applies regardless of which model actually serves a given request. See our Open-Source & Model Attributions.

3. Enforcement

We may investigate suspected violations and suspend or terminate access, remove content, or report unlawful activity. Serious or repeated violations may result in immediate termination without refund.

4. Reporting

Report abuse via Support & Data Rights.