TensorScale AI Acceptable Use Policy

TensorScale provides hosted model inference and related services. You are responsible for your use of the Services and for use by your Authorized Users and End Users. You will use the Services only in a lawful, safe, and responsible manner.

1. Illegal and harmful content

You will not use the Services to generate, upload, distribute, or facilitate:

• child sexual exploitation or abuse material, grooming, or any sexualization or exploitation of minors;
• non-consensual intimate imagery or deepfakes of a sexual nature;
• content that promotes or facilitates human trafficking, sexual violence, terrorism, or violent extremism;
• content that promotes self-harm or suicide;
• threats, harassment, bullying, or incitement of violence;
• hate speech or content that dehumanizes people based on protected characteristics;
• extreme gore or graphic violence intended to shock or harass; or
• any other content that is illegal under applicable law.

2. Fraud, deception, and abuse

You will not use the Services to:

• commit fraud, scams, phishing, credential theft, or social engineering;
• generate spam or unsolicited bulk communications;
• impersonate a person or entity in a commercial, defamatory, illegal, or harmful manner without authorization;
• create deceptive deepfakes or synthetic media intended to mislead about elections, public safety, or public health; or
• misrepresent AI-generated content as human-created where disclosure is required by law or the Documentation.

3. Privacy, IP, and third-party rights

You will not use the Services to:

• infringe copyrights, trademarks, publicity rights, privacy rights, or other third-party rights;
• submit personal data you are not authorized to process;
• submit protected health information or other regulated sensitive data unless TensorScale has agreed in writing that the Services are suitable; or
• collect or process biometric identifiers in violation of applicable law.

4. Security and platform integrity

You will not:

• probe, scan, or test the vulnerability of the Services without authorization;
• interfere with or disrupt the Services, other customers, or related networks;
• circumvent rate limits, authentication, billing, safety filters, watermarking, or provenance controls;
• attempt to extract model weights, non-public prompts, or proprietary systems except as allowed by an applicable open-source license or written permission; or
• use the Services to distribute malware or gain unauthorized access to systems.

5. Competitive and commercial misuse

You will not:

• use the Services to develop, train, or improve a competing inference platform or a direct substitute for the Services without TensorScale's prior written consent;
• resell or provide the Services as a standalone competing API or service bureau except through Customer Applications permitted by the Terms; or
• publish misleading benchmarks or disclose TensorScale Confidential Information.

6. High-Risk Uses

You will not use the Services for High-Risk Uses described in the Terms without TensorScale's prior written approval. That includes, among other things, uncontrolled medical, weapons, critical-infrastructure, and fully automated legal or credit decisions without meaningful human review, and prohibited AI practices under applicable law such as the EU AI Act.

7. Model licenses and sanctions

You will comply with all applicable Base Model and open-source licenses. You will not use the Services if you are under 18, located in a comprehensively sanctioned territory, or on a denied-party list, as further described in the Terms.

8. Customer Applications and End Users

If you offer Customer Applications to End Users, you must maintain terms and controls at least as protective as the Terms and this AUP, including age-gating and content moderation where required by law, and cooperate with TensorScale's reasonable enforcement requests.

9. Reporting and enforcement

Report suspected violations to legal@tensorscale.io. TensorScale may investigate, remove content, suspend or terminate accounts, block access, and report illegal activity to authorities.

TensorScale may update this AUP by posting a revised version. Continued use of the Services after the update constitutes acceptance. If this AUP conflicts with the Terms, the Terms control except that this AUP controls to the extent it imposes more restrictive use obligations.