Talk to a model on my own hardware.

This chat is served by a language model running on my self-hosted rig, streamed to you through a small rate-limited proxy I wrote. The same setup I deploy for clients who need data sovereignty.