Input-based LLM Routing Agent

🧐 I. Overview

A dynamic input-based LLM routing AI-agent built with n8n. It selects the most suitable language model:

• Sonar (Perplexity)
• GPT o3-mini-high (OpenAI)
• Claude 3.5 Sonnet (Anthropic)

based on user queries, and uses the selected model to answer, solve, or address the input accordingly.

⛓️ II. Workflow

The workflow above contains the following nodes:

• Chat Trigger (Chat Input)

• AI Agent (Routing Agent, Agent Output)

• OpenRouter Chat Model (gpt-4o-mini, Dynamic Model)

• Simple Memory (Memory 1, Memory 2)

• Structured Output Parser
  
  

🛠️ III. Use this repository

1. Make sure to have the following requirements:

• n8n account (Click here to create an account.)
• OpenRouter account + enough credits (Click here to create an account.)
• Docker Desktop (Optional for local hosting. Download here.)

2. Clone this repository.

Run this command in your terminal:

git clone https://github.com/m3mentomor1/Input-based-LLM-Routing-Agent.git

(Optional: You can also Fork this repository.)

3. Go to the repository's main directory.

Run this command in your terminal:

cd Input-based-LLM-Routing-Agent

4. Import workflow into n8n.

• Create a new workflow or open an existing one (n8n website or local via docker).
• Click the ••• menu in the top-right corner.
• Select Import from File...
• Choose workflow.json from this repository.