开发文档

开发文档

Acceptable Use Policy / 可接受使用政策

可接受使用政策_AUP.md

Acceptable Use Policy / 可接受使用政策

版本：v0.1 日期：2026-05-11 状态：模板，需律师和安全负责人审阅后上线

---

1. Purpose / 目的

This Acceptable Use Policy governs use of [Platform Name]. Users must comply with this policy, applicable laws, third-party model provider policies, and platform instructions.

We may suspend or terminate accounts, API keys, or traffic that violates this policy or creates legal, safety, provider, security, or platform risk.

---

2. Prohibited Activities / 禁止行为

You may not use the Services for:

Illegal activities:

  • Violating applicable law or regulation.
  • Circumventing sanctions, export controls, or restricted country rules.
  • Facilitating trafficking, exploitation, or organized crime.

Abuse and deception:

  • Phishing, credential theft, impersonation, or fraud.
  • Spam, unsolicited bulk messaging, or deceptive marketing.
  • Scams, fake reviews, fake engagement, or social manipulation.
  • Circumventing rate limits, safety systems, or access controls.

Cyber abuse:

  • Malware generation or deployment.
  • Credential stuffing.
  • Vulnerability exploitation without authorization.
  • Botnet, DDoS, or persistence tooling.
  • Instructions to compromise systems, accounts, or networks.

Privacy violations:

  • Doxxing.
  • Unauthorized surveillance.
  • Biometric identification or sensitive inference where unlawful or harmful.
  • Scraping, aggregating, or exposing personal data without lawful basis.

Harmful content:

  • Child sexual abuse material or exploitation.
  • Non-consensual sexual content.
  • Harassment, threats, or targeted abuse.
  • Instructions for self-harm or violence.
  • Extremist recruitment or operational support.

High-impact decisions without safeguards:

  • Credit, lending, insurance, housing, employment, education, legal, medical, or other high-impact decisions without human review, transparency, and compliance controls.

Model misuse:

  • Using outputs or interactions to train, extract, reverse engineer, or build competing models where prohibited by provider terms.
  • Model extraction, prompt extraction, or systematic scraping.
  • Misrepresenting model identity or output source.

---

3. Restricted Uses / 受限用途

The following uses require prior written approval or enterprise review:

  • Healthcare or medical workflows.
  • Legal or financial advisory workflows.
  • Government, law enforcement, or public-sector use.
  • Biometric, identity, or surveillance use.
  • Political campaigning, election-related persuasion, or lobbying.
  • High-volume messaging or customer outreach.
  • Use involving minors.
  • Sensitive personal data at scale.

---

4. User Responsibilities / 用户责任

Users must:

  • Maintain secure API keys.
  • Implement appropriate end-user terms and privacy notices.
  • Monitor their own applications for abuse.
  • Respect model provider restrictions.
  • Use human review for high-risk outputs.
  • Respond to platform inquiries about suspicious activity.
  • Report compromised keys or abuse.

---

5. Enforcement / 处置措施

We may take one or more actions:

  • Warning.
  • Rate limiting.
  • Model restriction.
  • API key suspension.
  • Account suspension.
  • Balance hold for fraud review.
  • Termination.
  • Reporting to providers, payment processors, or authorities where required.

We may consider intent, severity, volume, harm, repeat behavior, and cooperation.

---

6. Appeals / 申诉

Users may appeal enforcement decisions by contacting [Support Email] with:

  • Account email.
  • API key prefix.
  • Request IDs.
  • Explanation.
  • Remediation steps.

We are not required to restore access where doing so would create legal, provider, payment, security, or safety risk.

---

7. Provider Policies / 上游政策

Because the platform routes to third-party model providers, stricter provider-specific policies may apply. If a provider prohibits a use case, that use case is prohibited on our platform for that provider, even if not explicitly listed here.