Inferent

Content Policy

Version: 1.0Last Updated: July 2026Jurisdiction: Global

1. Overview

This Content Policy defines what types of content Leuks models may and may not generate. It applies to all users accessing Leuks models through Inferent's products or the Leuks API. This policy works in conjunction with our Model Usage Policy.

2. Safe Content

The following content is freely permitted:

  • Educational and informational content on virtually all subjects
  • Creative writing, fiction, poetry, scripts, and other creative works
  • Code and technical documentation
  • Business documents, summaries, and analysis
  • Translations and language assistance
  • Factual Q&A and research assistance
  • Discussion of sensitive historical events for educational purposes

3. Restricted Content

The following content types require appropriate context, purpose, or authorization, and are subject to additional review:

  • Adult content: Sexually explicit content is off by default and available only to verified adult platforms with an enterprise agreement
  • Detailed depictions of violence: Permitted in clearly fictional contexts; restricted when gratuitous or instructional
  • Medical and legal information: Permitted as general information with appropriate disclaimers; not as professional advice
  • Security research: General cybersecurity education is permitted; offensive exploit code requires authorization

4. Prohibited Content

The following content is absolutely prohibited and will never be generated:

  • Child sexual abuse material (CSAM) or any content that sexualizes minors
  • Instructions for creating weapons capable of mass casualties
  • Functional malware, ransomware, or cyberweapons
  • Content designed to facilitate human trafficking
  • Targeted harassment of specific individuals
  • Detailed instructions for self-harm or suicide directed at vulnerable individuals
  • Deepfakes of real individuals designed to cause harm or deceive without consent

5. AI-Generated Content Disclaimer

Content generated by Leuks models is produced by an AI system and may contain factual inaccuracies, outdated information, or reflect unintended biases. AI-generated content should not be relied upon as a sole source of truth for consequential decisions.

Developers who build products using Leuks must disclose to their end users when content is AI-generated, where such disclosure is required by law or where a reasonable user would want to know.

6. Human Review Policy

We maintain human review processes for flagged content and model safety incidents. High-severity flags—such as potential CSAM, imminent safety threats, or large-scale policy violations—are escalated for immediate human review. Our safety team reviews a sample of moderation decisions to ensure consistency and accuracy.

7. Prompt Injection Policy

Prompt injection is the practice of crafting inputs that attempt to override a model's instructions, bypass safety measures, or cause the model to act against its guidelines. Such attempts are a violation of our Terms of Service.

We invest in technical defenses against prompt injection and treat systematic attempts as security incidents subject to account suspension.

8. Enforcement & Appeals

Content that violates this policy may be blocked, removed, or result in account action. If you believe a content moderation decision was made in error, you may appeal via [email protected]. We review all appeals and aim to respond within 5 business days.