Six weeks of running a public Zerto Docs MCP server taught me the ceiling of dense-only retrieval. This is the story of finding it, fixing it with BM25 + RRF + reranker, and three other things (KB articles, a curated lessons doc, ingested ZertoPublic example scripts) that needed fixing while I was in there. The post Zerto Docs MCP — Part 2: Hybrid Search and a Curated Knowledge Layer appeared first on Justin's IT Blog.
No comments yet. Log in to reply on the Fediverse. Comments will appear here.