RAG Reranking

This system demonstrates a no-code AI agent that enhances vector-based document retrieval with a Cohere-powered re-ranker and metadata filtering. It uses a Superbase vector database to store chunked data and dynamically applies metadata filters for targeted queries. The workflow refines results by reranking returned chunks, ensuring accurate extraction of rule-based content. Developed by Nate Herk, the system is available via a free template in his Skool community.