Problem

Large collections of n8n workflows were difficult to search, analyze, and reuse efficiently. Manual browsing was slow and didn’t support semantic search.

Solution

• Built an automated AI knowledge base pipeline using n8n:
• Web scraping via HTTP requests, Headless Chrome (Browserless), and APIs
• Data normalization and pagination handling
• Storage in NocoDB for operational use
• Embedding generation and storage in Supabase vector DB
• Duplicate detection before database insertion
• AI agent connected via RAG architecture

Result

• Searchable AI-powered knowledge base
• Automated workflow ingestion pipeline
• Semantic search and AI-assisted retrieval

Skills:

1. n8n
2. Supabase Vector DB