Ante 0.2: a ~15MB coding agent that manages llama.cpp for you — point it at a GGUF and the whole agent loop runs offline

Hello~ We just shipped Ante 0.2, and the part I think this community will care about most is offline mode. We wanted local to be a first-class way to run the agent, so Ante manages the inference engine itself: Managed llama.cpp. Ante installs a pinned, checksum-verified official build matched to you