Do I need to install external dependencies?

Yes, this skill requires the Bun runtime and the qmd search engine to be installed on your local system to function.

What is the difference between keyword and semantic search?

Keyword search (BM25) finds exact word matches, while semantic search (vector) finds notes based on meaning and context, even if they don't share the same vocabulary.

When should I use hybrid search versus vector search?

Vector search is best for quick similarity lookups, while hybrid search (query) is recommended for thorough exploration as it combines multiple methods with LLM re-ranking.

How do I keep my search index up to date?

You can use the 'reindex' or 'rebuild index' commands within Claude to trigger 'qmd update', which scans for new, changed, or deleted files.

Does this skill send my notes to the cloud for searching?

No, this skill uses the qmd engine which performs all vector embeddings and searches locally on your machine, ensuring your data remains private.

Semantic Vault Search

Name: Semantic Vault Search
Author: JSai23

byJSai23

•

Productivity & Workflow

Enables deep semantic discovery and keyword searching within local Markdown vaults using vector-based and hybrid search algorithms.

This skill integrates the qmd local search engine into Claude, allowing users to perform sophisticated searches across their personal knowledge bases without relying on cloud APIs. It supports keyword-based BM25 search for specific terms, vector similarity search for finding conceptually related ideas, and advanced hybrid search with LLM re-ranking for complex queries. This is an essential tool for developers and researchers managing large Markdown vaults who need to find filing destinations, build backlinks, or uncover non-obvious connections between disparate notes.

Key Features

01Local vector similarity search for conceptual matching

02BM25 keyword search for precise term retrieval

03Support for multiple output formats including JSON and Markdown

04Automatic indexing and updating of Markdown collections

05Hybrid search with LLM re-ranking for maximum accuracy

061 GitHub stars

Use Cases

01Finding semantically related notes to build internal links and backlinks

02Discovering the optimal location to file new content based on existing vault context

03Identifying emergent patterns and connections across diverse subject areas

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jsai23/claude-tooling sem-search-a

For use in Claude.ai and ChatGPT