Video by NetApp Instaclustr via YouTube
Learn how Apache Cassandra 5.0 supports vector search for Retrieval-Augmented Generation (RAG) without a separate vector database. This demo covers embeddings, Storage-Attached Indexes (SAI), ANN queries, and three real-world retrieval scenarios, showing where vector search excels and where lexical search or metadata filtering is still needed.
Relevant links:
https://github.com/instaclustr/code-samples/tree/main/Cassandra/cassandra-vector-demo
https://www.instaclustr.com/blog/vector-search-apache-cassandra-rag-retrieval/
Timecodes
0:00 Introduction
0:28 The problem RAG solves
1:10 How Cassandra fits
2:37 Paraphrase query
3:07 Exact recall ID
3:49 Ambiguous query
4:10 When this architecture fits
4:57 Try it yourself