Terms, privacy, acceptable use.

Kitsune provides an OpenAI-compatible inference endpoint with zero data retention. Prompt and completion bodies are not stored. Request metadata (model, token counts, cost, latency and status) is kept for billing and the usage page. Access is provided as-is, without a guaranteed uptime SLA. Model availability and prices can change; the catalog and the price shown on each request are authoritative.

• The service is prepaid. You top up a balance in USD; each request is charged per token from that balance.
• There is no subscription, monthly minimum or unused allowance. Balance does not expire while the account is open.
• Cached prompt tokens are billed at the model cache price when one is published, otherwise at the input price.
• Sub-cent usage accrues and is charged in whole cents as it crosses the threshold.
• Failed calls are not billed. Aborted streams are billed for the usage reported by the service.
• Top-ups are credited after the payment is confirmed. Contact support for a billing correction before initiating a chargeback.

Kitsune stores account data (email, hashed password, sessions) and request metadata (model, token counts, cost, latency, status) for billing and the usage page. Prompt and completion bodies are not stored — zero data retention.

API keys are stored as SHA-256 hashes. Session cookies are HTTP-only. You can revoke keys and sessions from the console at any time.

Do not use Kitsune for illegal content, malware, fraud, harassment, or to evade applicable terms. Do not attempt to resell raw access without permission, probe the service for vulnerabilities, or use it to attack other systems. Accounts that abuse the service or generate fraudulent payments may be suspended and the remaining balance frozen pending review.

Billing, abuse or data requests: kitsune@kitsune.cv. Include the account email and, for billing, the payment id from the console.