Monitor AI Chat Interactions with Gemini 2.5 & Langfuse Tracing

This n8n workflow creates a simple yet powerful AI Agent for handling chat messages with Gemini 2.5, integrating Langfuse tracing to monitor LLM interactions in real-time. It uses a custom LangChain code node to initialize the LLM with callbacks, enabling seamless data export to Langfuse for detailed logs, performance metrics, and error tracking.

Key components include a chat trigger webhook, LangChain code node for LLM configuration (model, temperature, provider), and an AI Agent node that leverages the traced LLM. Setup requires self-hosted n8n (>=1.98.0), Langfuse credentials, and an LLM API key like Google for Gemini. Environment variables or direct node config handle authentication.

Benefits include enhanced observability for production AI chats, quick issue identification, cost optimization via usage insights, and scalability for agentic workflows. Ideal use cases: HR chatbots for payroll queries, customer support automation, internal ops tools, or any LLM-powered chat needing monitoring to ensure reliability and compliance.