Single-file, dependency-free live dashboard for local llama.cpp/vLLM serving — GPU, throughput, KV/ctx, model library. Stdlib Python + one local-first HTML page.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).