This n8n workflow template is designed to route user input to specialized agents (like a Reminder Agent, Email Agent, etc.) using a structured output from a language model. Here's a complete description of what it does and how each part works:
This template receives a user's request via Webhook, processes it using an LLM, extracts structured data like the agent name and user query, and routes the input to the appropriate sub-workflow (agent) based on the specified agent type.
Webhook
Node: GPT 4o Mini
Purpose: Interprets the user input and determines:
Auto-fixing Output Parser
Node: Structured Output Parser
Purpose: Converts the language model's response into a strict JSON structure with keys like:
"Agent Name"
"user input"
"sessionID"
Node: Switch ("Agent Route")
Purpose: Based on "Agent Name"
, it routes the input to one of the following sub-workflows:
Reminder Agent
Email Agent
Document Agent
Meeting Agent
Each agent is implemented as a separate n8n workflow:
"Agent Name"
is "Reminder Agent"
, the workflow "Reminder Agent"
is called with "user input"
.Respond to Webhook
node sends the output back as an HTTP response.User says: “Remind me to call my mom tomorrow.”
→ Routed to Reminder Agent
User says: “Send an email to the HR team.”
→ Routed to Email Agent
User says: “Schedule a meeting with John next week.”
→ Routed to Meeting Agent