LLM
Voice Agents use an LLM (Large Language Model) to understand conversational context and generate responses.
Lemu manages the underlying LLM for your voice agents. There is no per-agent provider or model picker to configure — Lemu selects and maintains the model powering your agents so conversations stay fast and reliable as models improve.
What you do control is the agent's behavior: the Global Prompt and each node's prompt, which together are sent to the LLM to generate every response. See Workflows and Agents for how prompts are structured.