Terms of Service
Last updated: May 2026. Please read carefully — using this site means you accept the terms below.
1. Account
- You must register with a real email address. Multiple accounts per person are not prohibited, but using multiple accounts to abuse quotas or evade risk controls is.
- You are responsible for safeguarding your account and API Keys. Losses caused by key leaks are your responsibility — delete and recreate immediately if leaked.
- Long-inactive accounts (12 consecutive months without any calls or top-ups) may be frozen. Frozen-account balances may still be withdrawn or migrated on request.
2. Acceptable use
- This site is an API gateway for AI models — we provide technical relay only and do not perform substantive review of the content models generate.
- The following uses are prohibited:
- Content that violates applicable laws and regulations
- Content that opposes constitutional principles, endangers national security, or harms national interests
- Content that incites ethnic hatred or discrimination, or undermines social cohesion
- Content that defames others or infringes on legal rights
- Content that is pornographic, violent, terror-related, or that incites criminal acts
- Large-scale unauthorized scraping, bot crawling, or attacks on third-party services
- Academic fraud, automated cheating on exams, or other violations of academic integrity
- When abnormal or violating calls are detected, this site may immediately suspend service, freeze the account balance, and cooperate with relevant authorities for investigation.
3. Billing and refunds
- Calls are billed per token usage at upstream model rates with a discount applied.
- Topped-up amounts are usable only on this site and are non-refundable in principle. In force-majeure cases (e.g. this site being shut down), remaining balances will be refunded by negotiation.
- Billing details are itemized in Console → Logs. For billing disputes, please reach out within 30 days of the call in question.
4. Privacy
- This site only logs metadata (timestamp, model, token, token usage, error code) — we do not persist request messages content or response bodies.
- Content you pass through is processed by the upstream model providers; their data-use policies apply (OpenAI, Anthropic, Google, Alibaba, ByteDance, etc.).
- For anti-fraud and incident-troubleshooting purposes, we may temporarily cache call metadata for ≤ 30 days. We do not share this with third parties.
5. Availability
- We make a best effort to keep the service available but do not guarantee 100% uptime. Planned maintenance is announced in advance on the main site or in the official group.
- This site is not liable for direct damages caused by upstream provider failures, network operator issues, or force majeure (natural disasters, policy changes, etc.). For production use, design for multi-model and multi-channel fallback.
6. Changes to these terms
We may revise these terms from time to time. Material revisions will be announced 7 days in advance via email or main-site announcement. Continued use constitutes acceptance of the revised terms.
7. Contact
For questions, complaints, reports, or compliance requests, please contact us via:
- Official group (QR code on the main site)
- Email:
support@aiproxynode.com