qmd

Semantic search of an Obsidian vault uses vector embeddings to match the meaning of your query, finding relevant notes even without exact keywords. This Skill combines vector search with BM25 keyword retrieval and LLM reranking for hybrid results.

Hybrid search for personal knowledge bases combines BM25 keyword search with vector semantic search and LLM reranking. This approach surfaces relevant notes, past decisions, and documentation by matching both exact terms and contextual meaning.

Yes, searching notes with AI requires the 'qmd' binary and an indexed Obsidian vault collection. The Skill depends on these components to execute keyword, semantic, and hybrid search queries against your personal documentation.

Finding past project decisions without exact keywords is possible through vector semantic search. The Skill interprets the meaning behind your query to locate relevant notes and documentation within your Obsidian vault.

The best way to recall personal preferences documented in Obsidian is using hybrid search, which combines fast BM25 keyword matching with vector semantic search and LLM reranking to surface accurate notes.

Semantic search requires an indexed vault collection because the 'qmd' binary must process and store vector embeddings of your notes beforehand. Without this index, the LLM reranking and vector retrieval cannot execute.