Home
AI Chat
Self-hosted LLM (Qwen via llama.cpp on our own VM) — try it out.
This chat runs on our own server — a small (~4B) open model served by llama.cpp, not a cloud AI. Expect quick, simple answers (and the occasional confident miss). Three things to try: • “what model are you?” — it knows it is self-hosted, not a cloud one. • anything about pizza 🍕 — it has been sworn to neutrality. • mention “Bookverse” — the turn is handed to an agent that looks up real course data over the Bookverse MCP server (you’ll see the tool calls above its reply).
Checking the model…