RAG Chatbot with OpenAI and Supabase Knowledge Base - n8n Workflow

Developers needing a powerful, stateful AI agent.
Businesses looking to implement a customer support chatbot based on internal knowledge.
n8n users interested in complex AI orchestration using the LangChain n8n node set.
Anyone wanting to leverage Supabase data within an n8n workflow for dynamic AI responses.

This sophisticated n8n workflow template solves the challenge of creating accurate, context-aware chatbots that can reference external data sources. Standard LLMs often hallucinate or lack access to proprietary data; this n8n workflow overcomes that limitation by implementing a Retrieval-Augmented Generation (RAG) pattern. The core of this system is the powerful Smart AI Agent n8n node, which uses the Supabase Database as a tool. When a user sends a query, the agent intelligently decides whether it needs to query the database before formulating a response using the connected OpenAI Chat Model. This structure ensures high relevance and accuracy, making this n8n workflow an essential tool for modern automation specialists.

The process begins with the Start Chat Conversation n8n trigger. This specialized n8n trigger waits for an incoming user message, typically from a connected chat interface.

1. Trigger and Orchestration: Once a message is received, the input is passed directly to the Smart AI Agent n8n node.


2. State Management: The agent immediately utilizes the Remember Chat History n8n node (a memory buffer) to maintain context across the conversation, ensuring the chatbot recalls past interactions.


3. LLM and Tooling: The agent is configured with the OpenAI Chat Model (for core reasoning and response generation) and the Supabase Database n8n node (configured as a tool).


4. Intelligent Retrieval: Based on the user's question, the agent determines if it needs information from the Supabase knowledge base. If necessary, it executes a query against the Supabase tool.


5. Final Response: The agent synthesizes the retrieved information (if any) with the current conversation history, and uses the OpenAI model to craft the final, relevant response. This complete n8n workflow structure provides a robust and functional RAG pipeline.

To use this powerful n8n workflow template, follow these steps:

1. Import: Copy the provided JSON code and import it into your self-hosted or cloud n8n instance using the 'New' -> 'Import from JSON' option.


2. OpenAI Credentials: Locate the OpenAI Chat Model n8n node and set up your OpenAI credentials (API Key). You may need to specify the model (e.g., GPT-4 or a fast GPT-3.5 variant).


3. Supabase Setup: Configure the Supabase Database n8n node. This requires credentials linking to your Supabase instance, including the URL and Service Role Key. Ensure the database table you intend to use as your knowledge base is set up correctly.


4. Agent Configuration: Review the Smart AI Agent settings, particularly the instructions and the names of the tools it can access.


5. Activation: Once credentials are set, activate the Start Chat Conversation n8n trigger by clicking 'Activate'. Your RAG chatbot n8n workflow is now ready to handle incoming queries.

Start Chat Conversation (n8n trigger): The initial n8n node that exposes a webhook endpoint, initiating the chat session when a message is received. It serves as the gateway for this specific n8n workflow.
Smart AI Agent: The central LangChain n8n node responsible for orchestration. It evaluates the user input, decides whether to consult its memory or its tools (the Supabase DB), and delegates tasks to the connected LLM.
OpenAI Chat Model: Provides the necessary GPT integration for high-quality language understanding and generation, connected directly to the Smart AI Agent as its core language model.
Remember Chat History: A LangChain memory n8n node that implements a buffer window memory, allowing the AI agent to reference previous messages in the current conversation, crucial for a stateful n8n workflow.

• Supabase Database: Configured as a tool within the n8n workflow. This standard n8n node allows the AI agent to dynamically retrieve structured or unstructured data from your Supabase instance, providing the knowledge base for RAG.