Search vault notes using BM25, vector search, and re-ranking.
npx skills add https://github.com/JSai23/claude-tooling --skill sem-search-a
Please help me install this Agent Skill. Skill: sem-search-a Source: https://github.com/JSai23/claude-tooling/tree/main/plugins/vault/skills/sem-search-a Command: npx skills add https://github.com/JSai23/claude-tooling --skill sem-search-a
Semantic search across vault notes is needed to surface conceptually related content when keyword matching fails, enabling faster discovery and better knowledge organization.
Provide a query text to search notes semantically, or run with --reindex to rebuild the index.
High-intent search queries and answers about installing and using this skill.
To perform semantic search across your notes vault, provide query text to find conceptually related content using BM25, vector search, and LLM re-ranking. This surfaces meaningfully related notes even when exact terms are forgotten.
Hybrid search combines BM25 and vector search by applying both techniques to vault notes and using LLM re-ranking to refine results. This method surfaces conceptually related content that keyword matching alone fails to find.
Yes, you need a local embeddings model and the qmd tool available in your environment to perform semantic search and generate the vector representations required for finding conceptually related notes.
The best way to find related notes when keyword matching fails is using local semantic search. It applies BM25, vector search, and re-ranking to locate conceptually related notes by querying meaning rather than exact terms.
To reindex your notes vault after adding new files, run the tool with the --reindex flag. This rebuilds embeddings and indexes to ensure semantic search results remain current with vault changes.
Semantic search returns outdated results from your notes vault when indexes are stale. Run with --reindex to rebuild embeddings and indexes after vault changes, keeping search results current.
JSai23Enable semantic search of local vault notes using qmd with BM25, vector search, and LLM re-ranking.
allenhutchisonSearch vault notes by meaning using semantic embeddings and RAG.
whmathews15Enable local semantic search for a personal vault using embeddings and SQLite.
JKhyroSearch local knowledge bases with hybrid keyword, vector, and reranked retrieval.
