Compile natural language specifications into tiny neural programs that run locally via llama.cpp. Once compiled, they work as regular Python functions.