**Auxen** provisions private AI model endpoints on dedicated GPU infrastructure. Customers pick a model from the catalog — open-source models like Llama 3.1, Qwen 2.5, Mistral, Gemma 2, and Phi-3 — and Auxen handles provisioning, scaling, billing, and teardown. Each instance is fully private: no shared inference, no third-party routing. **Why use it:** when you need real privacy (regulated industries — legal, healthcare, finance, government), high-volume inference where per-token APIs destroy unit economics, fine-tuning on your own data, or ephemeral compute for episodic agent workloads. Pay-per-minute billing — no subscriptions, no per-token fees. **Auth:** OAuth 2.1 + PKCE with Dynamic Client Registration ([RFC 7591](https://datatracker.ietf.org/doc/html/rfc7591)). Direct API keys (`auxen_live_*`) also supported for programmatic agents. **Tools (6):** - `auxen_list_models` — list available models (read-only) - `auxen_get_instance_status` — read instance state (read-only) - `auxen