Allowed
Coding agents, chat applications, automations, research, content workflows, structured output, and lawful API integrations are welcome. Normal high-volume use is subject to published fairness and capacity limits.
Not allowed
- Unlawful, fraudulent, deceptive, or rights-infringing activity.
- Malware, credential theft, phishing, unauthorized scanning or exploitation, and denial-of-service activity.
- Child sexual abuse material, sexual exploitation, or sexualization of minors.
- Instructions intended to cause serious physical harm, including weapons or explosives.
- Spam, mass manipulation, evasion of provider safety systems, or attempts to extract private models, prompts, keys, or infrastructure data.
- Automated mass account creation, shared-account resale, bypassing balances or rate limits, and abuse of promotions or payments.
Enforcement
We can throttle or suspend activity that creates security, legal, financial, or stability risks. We use metadata and automated safeguards, not routine prompt logging, to protect the service. Providers can independently refuse or terminate requests.
Reporting
Report abuse or security issues to [email protected]. Include request IDs and timestamps where useful, but never send full credentials.