Hi, I am using custom llm for my application. The responses are generated from nlp modules that would take some time so to keep it conversational I would like to add fillers and back channeling with emotional tone. How can I make best out of vapi?

Based on Vapi's documentation, I can see you're looking to enhance your CustomLLM with emotional speech elements. Vapi offers built-in features for backchanneling (using cues like "yeah," "uh-huh") and filler injection (adding "umm," "ahh," etc.). To better help you implement these features with your CustomLLM, could you share which specific TTS provider you're using with Vapi (ElevenLabs, etc.)? This would help me recommend the most appropriate approach for adding emotional elements to your responses. Sources: - [Vapi Documentation How Vapi Works](https://docs.vapi.ai/how-vapi-works) - [Vapi Documentation Custom LLM](https://docs.vapi.ai/customization/custom-llm/fine-tuned-openai-models) - [Vapi Documentation ElevenLabs Integration](https://docs.vapi.ai/providers/voice/elevenlabs)

Use features like backchanneling, emotion detection, and fillers, and integrate your LLM by setting up a local server, configuring Vapi, and crafting prompts that specify emotional tone and conversational cues.