API Configuration
OpenAI-compatible base URL (without
/v1/chat/completions)
Prints a reasoning note with each LLM call
Every function call is evaluated by an LLM
/v1/chat/completions)