Turn websites into RAG chatbot knowledge bases with Apify, OpenAI and Pinecone

This n8n template transforms any website into a fully functional RAG-ready chatbot knowledge base. It crawls sites using the AI Training Data Scraper community node, chunks content intelligently, generates embeddings, and stores everything in Pinecone for semantic search-powered conversations. Perfect for turning documentation, blogs, or marketing sites into instant AI chat assistants.

Use cases

• Convert documentation sites into intelligent support chatbots
• Build product knowledge bases from marketing websites
• Create internal search tools from company intranets
• Power customer support agents with scraped competitor analysis
• Generate training data for fine-tuning company-specific AI models

Good to know
This workflow connects to external services requiring API credentials. Works on n8n Cloud and self-hosted instances. Initial setup takes 10 minutes including community node installation.

Requirements

• n8n Cloud or self-hosted instance
• Community Node
• Apify API key
• OpenAI API key
• Pinecone account & index

Customising this workflow

• Replace Pinecone with Qdrant, Weaviate, or pgvector for self-hosted vector storage
• Swap OpenAI embeddings with Ollama or Hugging Face for zero-cost processing
• Add content filtering by language, code presence, or section type
• Extend with conversation memory using Redis or Postgres
• Build a frontend dashboard for managing multiple website indexes
• Add multi-site RAG (query across multiple domains simultaneously)