Custom AI Chatbot with Persona, Memory, and Language Model Integration

This n8n workflow enables the creation of a custom AI chatbot that interacts with users through a web interface, utilizing Google Gemini’s language model. The workflow is designed to set up an engaging persona, manage conversation history, and dynamically construct responses based on user inputs.

The process begins with a webhook trigger that activates whenever a chat message is received. It captures the message and stores the conversation history for contextual awareness. The core of the workflow is the ‘Construct & Execute LLM Prompt’ node, which creates a customized prompt acting as a persona — in this case, an eloquent and witty girlfriend who responds in Chinese with specific character traits and guidelines.

This prompt is sent to the Google Gemini Chat Model, which generates a contextually appropriate reply. The reply is then returned via the webhook, enabling real-time interactive conversations. Additional notes in the flow guide users on setting up API credentials, configuring the language model, and managing conversation history for optimal context longevity.

This workflow is highly useful for developing personalized virtual assistants, conversational agents, or AI-driven chat companions with distinct personalities for marketing, customer service, or entertainment purposes.