How can I create an n8n chatbot that uses my website as its knowledge base?

Based on our documentation, you can achieve this by creating a separate workflow that serves as your website data source and then connecting it as a tool to your AI agent. Although there's no out-of-the-box tutorial specifically for extracting data from a website’s content, the recommended approach is similar to how you might import custom data into your agent.

Here’s a high-level approach:

1. Create a Data-Fetching Workflow:
   Build a workflow whose task is to fetch and process data from your website. This could involve using HTTP Request nodes or other appropriate methods to scrape or retrieve content from your site's API or endpoints. You can also add processing logic to filter or structure the data as needed.

2. Use the Custom Workflow Tool Node:
   Once your data-fetching workflow is set up, use the Custom n8n Workflow Tool node to connect this workflow to your AI Agent. This node lets you call your custom workflow as a "tool" from the agent. In the tutorial “Build an AI workflow in n8n,” a similar method is used to pull in custom data (the example uses return data about fruits).

3. Integrate with Your AI Agent:
   In your main AI workflow, add the Custom Workflow Tool node as one of the tools connected to your AI Agent. Configure its parameters (workflow ID, description, etc.), so that whenever the agent needs to look up information from your website, it calls this workflow. You can also pass parameters—for example, filtering criteria—to your workflow tool node, ensuring that it returns only the relevant content.

4. Test the Integration:
   Once configured, test your chatbot by sending queries that prompt the AI Agent to retrieve and process information from your website data source. Adjust the fetching and processing logic as needed to improve the results.

Additional Tips:

• If you haven’t already, review the tutorial on building AI workflows using n8n’s LangChain functionality. It walks through adding a chat trigger, connecting with OpenAI Chat Model, and integrating custom workflows.
• Make sure your website data-fetching workflow is robust, handling errors or changes in site structure, since reliable data retrieval is key for accurate responses.

This strategy lets your chatbot use your website as a dynamic knowledge source. While there isn’t a dedicated tutorial solely for website integration, leveraging the existing example for custom data integration should give you a strong starting point for your project.