1. Normal permitted use
AI training and inference, rendering, development, websites and APIs, research computing, data processing, backups, and other lawful business are generally permitted, subject to the order, resource limits, and upstream facility rules.
2. Prohibited conduct
Do not deploy malware, phishing, fraud, credential theft, DDoS or attacks, unauthorized scanning, spam, unlawful content, copyright or trademark infringement, unauthorized third-party mining, or attempts to bypass access, payment, or platform security controls.
3. High-risk workloads
Large-scale crawling, proxy egress, open relays, sustained scanning, unusual bandwidth use, and workloads that may affect neighboring networks must be disclosed and approved before deployment. Without approval, we may limit traffic, suspend the instance, or require a migration plan.
4. Response
When a credible report or clear risk is found, we may request information, limit resources, suspend service, or terminate an order, retaining necessary records where lawful. An urgent security event may be handled before notice.
Last reviewed: 2026-11-14. Product, price, inventory, and payment facts remain subject to the current order confirmation.