Custom AI Agent with LangChain & Gemini (Self-Hosted)

This workflow implements a fully customizable conversational AI agent leveraging the LangChain Code node integrated with Google Gemini API, designed exclusively for self-hosted n8n instances. It provides granular control over agent prompts, personality, and conversation structure, avoiding unnecessary token consumption from built-in tool-calling features in n8n's standard agents. Users can test directly in the editor or activate a web-based chat interface via the 'When Chat Message Received' trigger node.

Key benefits include flexible model swapping (e.g., other LLMs), adjustable conversation history length for memory management, and customizable UI elements like chat titles. The prompt template must retain {chat_history} and {input} placeholders for seamless LangChain operation, enabling advanced prompt engineering tailored to specific use cases such as customer support bots, personal assistants, or interactive Q&A systems.

Ideal for developers and businesses needing cost-effective, high-control AI chat solutions. Setup requires Gemini API credentials and self-hosted n8n; customization options span interface, prompts, models, and memory, making it versatile for personal projects or scalable applications while minimizing API costs through optimized token usage.