Model settings define the intelligence layer of your AI agent. Here you can select the LLM provider, choose a specific model, and configure a global prompt to guide the agent’s behavior.

​
Provider Selection

You can choose from multiple AI providers directly in the dropdown:
  • OpenAI (GPT-4, GPT-4 Mini, GPT-4 Nano)
This allows you to decide where the agent’s reasoning and language generation comes from.
Choosing the right provider balances between accuracy, latency, and cost.

​
Model Selection

Within each provider, you can select the specific model variant that best suits your use case. Examples for OpenAI:
  • GPT-4 → High intelligence, best for complex conversations
  • GPT-4 Mini → Lightweight, faster responses, lower cost
  • GPT-4 Nano → Ultra-fast for simple workflows
Agent Speech Settings
Other providers offer similar tiered options (high-quality vs. low-latency).

​
Global Prompt

The Global Prompt defines the personality and behavior of your AI agent across all conversations. It answers questions like:
  • How should the agent greet customers?
  • Should it behave as formal, friendly, professional, or casual?
  • What rules or tone should the agent always follow?
Agent Speech Settings Example: 
You are SigmaMind AI, a helpful virtual assistant. 
Always greet customers politely, answer their questions in simple language, 
and provide solutions in less than 3 steps when possible.
The global prompt applies to all nodes and workflows.
Use it to enforce a consistent brand voice and response style.

​
Summary

  • Select a provider (OpenAI, Cartesia, RhymeAI, or custom).
  • Pick a model based on latency, cost, and intelligence needs.
  • Define a global prompt to shape the agent’s behavior and tone.
Model settings are the brain of your agent, ensuring it thinks and responds the way you need.