Privacy policy · July 19, 2026

Your data, clearly explained.

What we keep

The testing log stores the raw ModelKlub API key used, complete request body, prompts, system instructions, messages, tool definitions, selected model, generated response, response headers, token totals, timestamps, request duration, success or failure metadata, and anti-abuse information such as a hashed IP address and country code. Payment records remain limited to the confirmation data needed to provide your plan.

Why we process this data

To test and improve gateway compatibility, investigate failed or abusive requests, authenticate access, apply token quotas, prevent key sharing, and provide support. Upstream AI providers also process request content under their own terms and privacy policies.

Storage and access

Captured inference content, full ModelKlub API keys, linked Discord IDs, request headers, IP data, matches, and responses are currently stored in plaintext at rest in a separate SQLite database on the Oracle host. The logger has no public network port. A second plaintext audit copy is posted to a private Discord channel restricted to the Owner role and the publishing bot. Upstream provider credentials are not exposed to customers or written into this log by ModelKlub.

Retention and control

Full captured inference logs in the Oracle testing database are automatically deleted after up to six hours. Owner-only Discord audit copies remain until an Owner deletes the messages or channel. Usage metadata may be retained longer where operational, billing, security, or legal needs require. You can revoke an API key at any time. For account deletion requests, use the support panel in the ModelKlub Discord.

What we do not use it for

We do not sell logged content, use it for advertising, or train AI models on it. The website does not use advertising trackers.

Contact

Questions or deletion requests: ModelKlub Discord.