Dynamic AI model router for query optimization with OpenRouter

The Agent Decisioner is a dynamic, AI-powered routing system that automatically selects the most appropriate large language model (LLM) to respond to a user's query based on the query’s content and purpose.

This workflow ensures dynamic, optimized AI responses by intelligently routing queries to the best-suited model.

Advantages

• 🔁 Automatic Model Routing: Automatically selects the best model for the job, improving efficiency and relevance of responses.

• 🎯 Optimized Use of Resources: Avoids overuse of expensive models like GPT-4 by routing simpler queries to lightweight models.

• 📚 Model-Aware Reasoning: Uses detailed metadata about model capabilities (e.g., reasoning, coding, web search) for intelligent selection.

• 📥 Modular and Extendable: Easy to integrate with other tools or expand by adding more models or custom decision logic.

• 👨‍💻 Ideal for RAG and Multi-Agent Systems: Can serve as the brain behind more complex agent frameworks or Retrieval-Augmented Generation pipelines.

How It Works

1. Chat Trigger: The workflow starts when a user sends a message, triggering the Routing Agent.
2. Model Selection: The AI Agent analyzes the query and selects the best-suited model from the available options (e.g., Claude 3.7 Sonnet for coding, Perplexity/Sonar for web searches, GPT-4o Mini for reasoning).
3. Structured Output: The agent returns a JSON response with the user’s prompt and the chosen model.
4. Execution: The selected model processes the query and generates a response, ensuring optimal performance for the task.

Set Up Steps

1. Configure Nodes:

   • Chat Trigger: Set up the webhook to receive user messages.
   • Routing Agent (AI Agent): Define the system message with model strengths and JSON output rules.
   • OpenRouter Chat Model: Connect to OpenRouter for model access.
   • Structured Output Parser: Ensure it validates the JSON response format (prompt + model).
   • Execution Agent (AI Agent1): Configure it to forward the prompt to the selected model.

2. Connect Nodes:

   • Link the Chat Trigger to the Routing Agent.
   • Connect the OpenRouter Chat Model and Output Parser to the Routing Agent.
   • Route the parsed JSON to the Execution Agent, which uses the chosen model via OpenRouter Chat Model1.

3. Credentials:

   • Ensure OpenRouter API credentials are correctly set for both chat model nodes.

4. Test & Deploy:

   • Activate the workflow and test with sample queries to verify model selection logic.
   • Adjust the routing rules if needed for better accuracy.

Need help customizing?

Contact me for consulting and support or add me on Linkedin.