Acceptable Use Policy
The rules for using the Inferway Service responsibly, how we enforce them, and how Zero Data Retention shapes abuse investigations.
GWMM LLC · a company registered in Wyoming, United States at 30 N Gould St, Sheridan, WY 82801
Acceptable Use Policy
This policy sets the rules for using the Inferway Service responsibly, explains how we enforce it, and describes how our zero-data-retention design shapes abuse investigations.
1. Scope
This Acceptable Use Policy ("AUP") applies to all use of the Inferway Service, including the API, web Sandbox, console, and website. It is part of the Terms of Service. Capitalized terms not defined here have the meaning given in the Terms.
2. Prohibited Uses
You may not use the Service, and may not allow others to use the Service, to:
- Violate the license terms or prohibited-use policy of the underlying model
- Generate, distribute, or facilitate unlawful, harmful, harassing, defamatory, infringing, or hateful content
- Generate, distribute, or solicit child sexual abuse material (CSAM) or any sexual content involving minors
- Generate, distribute, or solicit sexually explicit or intimate imagery of a real, identifiable person without their consent, including face-swapped or otherwise synthetic imagery
- Develop, promote, or deploy malware, exploits, ransomware, phishing, or other tools intended to harm or compromise systems or individuals
- Attempt to evade economic sanctions, export controls, or access the Service from prohibited jurisdictions
- Bypass rate limits, authentication, safety filters, or other technical protections, including by creating multiple accounts or API keys to circumvent restrictions
- Conduct load tests, stress probes, or other activity that exceeds the published concurrency ceiling or otherwise degrades the Service for others
- Resell or redistribute access to the Service in breach of the relaying and resale conditions in our Terms of Service, including by running an open, unauthenticated proxy or reselling promotional, free-tier, or gift credit
- Scrape, crawl, or reverse-engineer the Service or its underlying models
- Use the Service in any way that violates applicable law or the rights of any third party
3. Enforcement
We enforce this AUP using a graduated response that matches the severity and persistence of the violation.
- Rate limiting. Automated throttling for traffic patterns that exceed published limits or appear abusive.
- Suspension. Temporary suspension of an API key or account while we investigate suspected violations.
- Termination. Permanent termination of access for serious or repeated violations, including any violation of the underlying model's license terms referenced above.
We may act without prior notice where necessary to protect the Service, comply with law, or prevent harm.
4. ZDR & Investigations
Ordinary self-hosted text inference processes prompts and completions in memory without storing or logging them; the result-file, input-media, H3 prompt and optional-processing exceptions in the Privacy Policy and DPA also apply. This has important consequences for how we investigate abuse:
- Ordinary text inference has no historical content to provide. H3 prompts, rewrites and MiMo drafts not yet erased, and media not yet deleted, may exist under the disclosed exceptions; lawful preservation and disclosure follow applicable procedures. Database ciphertext is erased after no more than 7 days; backups rotate out within 30 days and are outside database erasure.
- We enforce this policy based on metadata patterns. We ordinarily detect abuse by examining signals such as request rate, volume, error patterns, and source behavior, not by reading the content of your prompts or completions.
- We will cooperate with lawful process. Where required by law, we will provide the non-content metadata we retain. We report suspected CSAM to the appropriate authorities as required by law.
Historical ordinary text inference content cannot be recovered for an investigation. MiniMax official enhancement and future 2K upgrades are customer-selected and separately consented; MiniMax retention and training terms remain unverified.
5. Reporting Abuse
If you become aware of abuse of the Service, please report it to hello@inferway.ai with as much detail as possible, including the time, affected endpoint, and the nature of the concern. While we cannot disclose the content of other users' requests, we will investigate metadata patterns and take action as appropriate.
6. Changes
We may update this AUP from time to time. Material changes will be posted on this page with at least 30 days' notice, unless law or an active threat requires earlier action. Continued use of the Service after the effective date constitutes acceptance.
7. Contact
Questions about this policy? Reach us at: