Apex Inference — Terms

Terms of Service

Zero-retention. Prompts and completions are processed exclusively in volatile GPU memory and immediately discarded. No logging or training.

1. The service

Apex Inference provides an API for running large language models on dedicated GPU capacity. Access is granted via API key. Capacity is provisioned as isolated pools; you do not share compute or memory with other customers.

2. Use

You agree to use the service in compliance with applicable law and to keep your API keys confidential. You are responsible for activity under your keys. You may not resell access except under a separately executed agreement.

3. Acceptable content

You may not use the service to generate or distribute content that violates law, or to run workloads intended to harm, disrupt, or compromise systems. We may suspend access to protect the service or other customers.

4. Data and retention

Prompts and completions are processed exclusively in volatile GPU memory and immediately discarded. We do not log request bodies and we do not train on your traffic. Aggregate, non-content usage telemetry may be retained for billing and operations.

5. Billing

Usage is billed per token, per the rate agreed for your pool, unless a volume contract applies. Invoices are due net 30. Overdue accounts may be suspended after notice.

6. Availability

We target the service levels stated in your agreement, but we do not guarantee uninterrupted availability. Scheduled maintenance and emergencies may cause brief interruptions. We are not liable for indirect or consequential damages arising from use of the service.

7. Termination

Either party may terminate the agreement on written notice. On termination, outstanding balances remain due, and your keys are revoked.

8. Contact

Questions or disputes: [email protected]. We will respond within five business days.

Last updated: 2026-08-18 · Privacy Policy