AI Orchestrator: Dynamic LLM Selection by Input Type

This n8n workflow intelligently classifies incoming chat messages using GPT-4.1-mini and routes them to specialized LLMs like GPT-4.1 mini or Gemini Flash, optimizing for speed, cost, and performance. It handles inputs via a 'When chat message received' trigger, including chatInput and sessionId for personalized sessions.

Key benefits include smart model routing to reduce expenses on heavy models, scalability for adding new types or LLMs, clear logic separation for easy maintenance, and support for per-user memory. Ideal for AI chatbots, virtual assistants, or customer support systems needing dynamic AI responses.

The process flows from classification into categories—general, reasoning, coding, search—via structured output parsing, followed by model selection and execution. Use cases span conversational AI in apps, enterprise chat agents, or IoT dashboards requiring adaptive intelligence, saving time on manual model management.