n8n templatesn8n templates

RAG & GenAI App With WordPress Content

This workflow automates the extraction of article and page content from WordPress websites to create an intelligent question-and-answer system based on retrieval-augmented generative artificial intelligence. It filters, transforms, and vectorizes the content, storing the data in a Supabase database to support efficient semantic retrieval and dynamic questioning. By integrating OpenAI's GPT-4 model, users can enjoy a more precise query experience while achieving persistent management of chat memory, enhancing the contextual continuity of interactions and increasing the intelligent utilization value of the website's content.

Workflow Diagram
RAG & GenAI App With WordPress Content Workflow diagram

Workflow Name

RAG & GenAI App With WordPress Content

Key Features and Highlights

This workflow implements a Retrieval-Augmented Generation (RAG) application based on WordPress website content. It automatically crawls articles and pages from WordPress, performs content filtering, transformation, and vectorization, then stores the embedding data in the Supabase vector database to support intelligent content-based Q&A. By integrating the OpenAI GPT-4 model with Postgres for persistent chat memory, it enables dynamic updates and intelligent interactions with website content.

Core Problems Addressed

  • Automates synchronization of the latest content from WordPress sites, eliminating manual data updates.
  • Utilizes vector embedding technology for efficient semantic search across large-scale website content.
  • Overcomes limitations of traditional Q&A systems that cannot accurately leverage rich website content, enhancing user query experience.
  • Supports filtering by publication status and content protection to ensure sensitive information is not exposed.
  • Implements persistent chat history management to improve contextual continuity in user interactions.

Application Scenarios

  • Enterprises or personal websites aiming to enhance user experience and content utilization through intelligent Q&A bots.
  • Content-rich WordPress sites such as blogs and news portals building content-driven intelligent customer service or assistants.
  • Use cases requiring regular synchronization of website content and knowledge base construction.
  • Integration with Generative AI technologies for intelligent content recommendation, Q&A, and interaction.

Main Workflow Steps

  1. Trigger and Content Acquisition
    • Workflow can be triggered manually or on a schedule.
    • Uses WordPress nodes to fetch all posts and pages, supporting incremental retrieval based on last execution time.
  2. Data Filtering and Processing
    • Filters to retain only published and unprotected content.
    • Converts HTML content to Markdown format for easier downstream processing.
  3. Content Splitting and Embedding Generation
    • Splits content into specified-sized chunks using a Token Splitter while maintaining contextual continuity.
    • Generates content vectors using OpenAI’s text embedding model (text-embedding-3-small).
  4. Vector Data Storage and Update
    • Checks if content already exists in the Supabase vector database.
    • Deletes and reinserts updated content; inserts new content directly.
    • Logs workflow execution history to facilitate incremental updates.
  5. Intelligent Q&A Interaction
    • Listens for chat triggers and receives user input.
    • Retrieves relevant documents from Supabase based on user queries.
    • Combines OpenAI GPT-4 model with Postgres chat memory to generate responses enriched with document metadata.
    • Responds to user requests via Webhook nodes, completing the Q&A interaction.

Systems and Services Involved

  • WordPress: Source of website articles and page content.
  • OpenAI: Text embedding generation (text-embedding-3-small model) and conversational generation (gpt-4o-mini model).
  • Supabase: Storage and management of vector data and workflow execution history.
  • Postgres: Persistent storage of chat history supporting vector retrieval.
  • n8n Core Nodes: Manual and scheduled triggers, HTTP requests, Webhook responses, data processing (aggregation, filtering, field setting, etc.).

Target Users and Value Proposition

  • Website operators and content managers seeking to enhance intelligent utilization of content through automation tools.
  • Developers and technical teams looking to rapidly build intelligent Q&A systems based on WordPress content.
  • Enterprise customer service teams aiming to create intelligent chatbots integrated with website content.
  • Organizations wanting to combine generative AI and retrieval-augmented techniques to improve user interaction quality and content search experience.

This workflow achieves a highly automated, multi-system integrated closed loop from WordPress content to intelligent Q&A, significantly enhancing the intelligent application value of website content.