Acceptable Use Policy

1. Core Principles

ClaudeKeys is committed to providing developers, researchers, and enterprises with frictionless, programmatic access to state-of-the-art AI inference. While we champion developer autonomy and privacy, we maintain a firm commitment to preventing malicious, harmful, or illegal misuse of computational resources.

This Acceptable Use Policy ("AUP") defines clear operational boundaries for all applications, automated agents, scripts, and users utilizing our API endpoints.

2. Strictly Prohibited Use Cases

• Child Sexual Exploitation and Abuse (CSAM): Any generation, depiction, facilitation, or discussion promoting sexual abuse or exploitation of minors.
• Violent Extremism & Terrorism: Content that plans, promotes, incites, or instructs acts of terrorism or severe physical violence against persons or critical infrastructure.
• Suicide and Self-Harm: Instructions, encouragement, or assistance for self-harm or suicide.
• Non-Consensual Sexual Content: Generating deepfake pornography, explicit imagery without consent, or targeted sexual harassment.
• Weapons Proliferation: Designing, synthesizing, or manufacturing chemical, biological, radiological, or nuclear (CBRN) weapons or high-yield explosives.

3. Cybersecurity & Malware Restrictions

While defensive security research, authorized penetration testing, and vulnerability remediation are permitted, using ClaudeKeys to facilitate offensive cyber operations is prohibited:

• Generating polymorphic malware, ransomware, weaponized exploits, or rootkits;
• Developing automated vulnerability scanner payloads intended for unauthorized network intrusion;
• Automating credential stuffing, brute-force dictionaries, or spear-phishing attack campaigns;
• Command-and-control (C2) communication automation for botnets.

4. Infrastructure Integrity & Rate Limits

To preserve low latency and high reliability across our 33 model gateways, developers must respect architectural boundaries:

• No Distributed Denial of Service (DDoS): Flooding endpoints beyond allocated plan rate limits (RPM/TPM) using distributed botnets or forged headers is prohibited.
• Key Pooling & Evasion: Creating multiple free accounts to circumvent subscription rate limits or token allowances is strictly monitored and will result in IP pool bans.
• Reverse Engineering: Attempting to probe or reverse-engineer ClaudeKeys internal proxy routing tables or credential stores is forbidden.

5. Alignment with Upstream Model Provider Policies

Because ClaudeKeys aggregates models developed by leading AI labs, developers are also subject to the foundational terms of each respective provider:

• OpenAI Usage Policies (when calling GPT-5.6, GPT-5.5, etc.);
• Anthropic Commercial Terms (when calling Claude Sonnet 5, Claude Opus 5, etc.);
• Google Generative AI Acceptable Use Policy (when calling Gemini 3.7 Flash, Gemini 3.1 Pro);
• DeepSeek, Meta LLaMA / Muse, Alibaba Qwen, and Zhipu GLM acceptable use guidelines.

6. Violations & Enforcement

ClaudeKeys employs automated heuristic anomaly detection to identify malicious traffic patterns (e.g. rapid 4xx error loops, DDoS signatures) without inspecting plaintext prompt payloads.

Upon detection of an AUP violation:

• We reserve the right to immediately suspend or permanently terminate offending API keys;
• Any unspent cryptocurrency credit balance associated with the account is forfeited as liquidated damages;
• Where required by mandatory criminal statutes (such as CSAM), we promptly submit necessary reports to the National Center for Missing & Exploited Children (NCMEC) or relevant international law enforcement bodies.

7. Reporting Abuse

If you suspect that an application or entity is utilizing ClaudeKeys to violate this policy, please submit an abuse notification to abuse@claudekeys.store with relevant technical headers and details.