Multi-Agent AI Conversation Workflow

This n8n workflow facilitates a scalable multi-agent AI conversation system, enabling users to interact with multiple AI assistants simultaneously within a chat interface. It dynamically manages interactions by parsing user input for @mentions, triggering relevant AI agents, and consolidating their responses for a seamless multi-agent dialogue. The workflow begins with a webhook trigger that captures incoming chat messages. It then defines global and agent-specific settings, configuring different AI models and system prompts for each agent. Extracted mentions from user input determine which agents are invoked. If no mentions are found, all agents are engaged in a randomized order. Each agent’s response is generated using the OpenRouter chat model, stored in a shared memory buffer for context-aware multi-turn conversations, and finally, all responses are combined and formatted into a single cohesive output. This workflow is ideal for scenarios like customer support, complex decision-making, or collaborative brainstorming, where multiple AI perspectives enhance the interaction experience.