This document is provided in English only.

LLM Acceptable Use Agreement

Last updated: 2026-04-25

This LLM Acceptable Use Agreement (“Agreement”) applies to SinoFlare large language model APIs and related offerings (collectively, “LLM Services”), including P1 hosted API access and any self-hosted or dedicated plans (such as P2). You must read and accept this Agreement before ordering or using LLM Services.

Mainland China prohibition. LLM Services are not available to users located in, residing in, or ordinarily operating from the Mainland of the People’s Republic of China (“Mainland China”). You represent and warrant that you are not a Mainland China user, that you will not access LLM Services from Mainland China, and that you will not provide LLM credentials, proxies, or outputs for use by Mainland China users. Accounts that select Mainland China as their region, or that we reasonably believe are used from or for Mainland China, are blocked from LLM Services and may be suspended without refund.

1. Scope

LLM Services may include model inference, token quota, API keys, upstream routing, and optional dedicated deployment. Features, models, pricing, peak/valley billing, and availability may change. Upstream providers’ terms also apply to traffic we forward on your behalf.

2. Eligibility and geo restrictions

In addition to the Mainland China prohibition above, we may restrict LLM Services in other jurisdictions to comply with law, sanctions, export controls, or provider requirements. You must keep your profile region accurate. Circumvention (VPNs used to evade geo policy, shared keys, false region selection) is a material breach.

3. Prohibited uses

You must not use LLM Services to:

Violate applicable law, including export control, sanctions, privacy, or intellectual property law;
Create or disseminate child sexual abuse material, non-consensual intimate imagery, or content that exploits minors;
Conduct scams, phishing, malware distribution, unauthorized access to systems, or other cybercrime;
Engage in violent extremism, terrorism, or credible threats of violence;
Build biometric surveillance systems for mass targeting without lawful basis;
Generate spam at scale, deceptive deepfakes for fraud, or medical/legal/financial advice presented as professional advice without human review;
Attempt to extract model weights, bypass safety filters, or overload infrastructure.

4. Your content and outputs

You are solely responsible for prompts, uploaded data, and how you use model outputs. Outputs may be inaccurate or inappropriate; you must independently verify results before relying on them for consequential decisions. Do not submit secrets you are not authorized to process.

5. Keys, quotas, and security

API keys and credentials are confidential. You are responsible for usage under your keys, including team sharing you enable. Quota, rate limits, and validity windows are enforced as described in the console. Abuse may result in immediate key revocation.

6. Fees and prepaid quota

LLM Services are prepaid unless we agree otherwise in writing. Payment-channel fees, if any, are disclosed at checkout. Unused quota generally expires according to the purchased validity period and is non-refundable except where required by law.

7. Monitoring and enforcement

We and upstream providers may log metadata and limited content signals for security, billing, and abuse prevention. We may investigate suspected violations, require additional verification, suspend LLM access, or terminate accounts. Severe violations may be reported to authorities or providers.

8. No professional advice; disclaimer

LLM OUTPUTS ARE NOT A SUBSTITUTE FOR PROFESSIONAL ADVICE. LLM SERVICES ARE PROVIDED “AS IS”. TO THE MAXIMUM EXTENT PERMITTED BY LAW, WE DISCLAIM LIABILITY FOR DECISIONS MADE BASED ON MODEL OUTPUTS AND FOR UPSTREAM OUTAGES BEYOND OUR REASONABLE CONTROL.

9. Changes

We may update this Agreement by posting a new version. Continued use of LLM Services after the effective date constitutes acceptance. If you do not agree, stop using LLM Services.

10. Acknowledgement

By accepting this Agreement in the console, you confirm that you have read it in English, that you understand the Mainland China prohibition, and that you will comply with all restrictions stated herein.