Supabase AI Agent Server with RAG & Multi-Tenant CRUD

This n8n workflow creates a robust Supabase-powered server for AI agents, integrating Retrieval-Augmented Generation (RAG) with OpenAI embeddings and Supabase vector search. It enables context-aware responses by retrieving relevant knowledge from vector stores, while supporting full CRUD operations on tables like agent_messages, agent_tasks, agent_status, and agent_knowledge. Multi-tenancy is achieved through dynamic table names and webhook paths, ensuring data isolation per user or organization, secured by Supabase Row Level Security (RLS).

Key benefits include persistent conversation history for chatbots, real-time task tracking, and scalable knowledge management without custom backend development. It saves significant time by automating AI-driven data persistence and retrieval, reducing reliance on stateless APIs. The workflow is production-ready for n8n version 1.88.0+, with MIT licensing for easy customization.

Ideal use cases: Customer support chatbots that remember past interactions and pull from institutional knowledge; automated task orchestration with status updates; and domain-specific knowledge repositories for AI agents. Setup involves importing the template, configuring Supabase and OpenAI credentials, optionally enabling multi-tenancy via dynamic paths, and activating webhooks for seamless integration.