Local llama.cpp model hot-swap controller for preserving chat context with low VRAM overhead
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).