Hey guys, Im using CustomLLM because I obviously want to manage all text-side-of-things nad reasoning on my LLM and use Vapi to transcribe-tts-call. What the hell do I send to vapi when in calls my endpoint exactly? I want to send text that will be TTSed wo whomever Im calling obviously. These did not help, vapi bot in support sections just halucinates absolute nonsense. https://docs.vapi.ai/customization/custom-llm/fine-tuned-openai-models#using-your-server https://github.com/VapiAI/server-example-python-flask/blob/main/app/api/custom_llm.py