docs(wiki): weighted RRF tuning and BM25 postings latency numbers
dcbc976
docs(wiki): update guides for reranking cost, multi-document, and pgvector schema
Note the reranking call in the cost and latency tables, describe the
multi-document upload and citation UI in Quick-Start, bump the Docker base image
to Node 24, and rewrite Swap-to-pgvector around the new chunk shape (id, docId,
page, source) with full-text search keeping the lexical half of hybrid search.
53943a0
docs(wiki): rewrite core pages for hybrid search, reranking, citations, multi-document
Rewrite Home, Architecture, How-RAG-Works, and Roadmap to describe the 1.1
pipeline: page-by-page parsing, hybrid retrieval with RRF fusion, the reranker
stage, NDJSON citation streaming, multi-document chat, and page-level
highlights. Update diagrams, the component table, and the shipped list.
933b6c6
docs(wiki): add architecture overview, examples, and troubleshooting to Home
ff9845a
docs(wiki): write Home — overview, who it's for, wiki index
86140d8
Wiki: Home, Architecture, How RAG Works, Quick Start, Configuration, Swap to pgvector, Cost and Performance, Deployment, Roadmap
44133cb