Search local knowledge bases with hybrid BM25, vector, and LLM reranking.
npx skills add https://github.com/valentinuuiuiu/vikarma --skill qmd-valentinuuiuiu
Please help me install this Agent Skill. Skill: qmd Source: https://github.com/valentinuuiuiu/vikarma/tree/main/hermes_agent/optional-skills/research/qmd Command: npx skills add https://github.com/valentinuuiuiu/vikarma --skill qmd-valentinuuiuiu
This Skill addresses the challenge of finding and retrieving information from personal knowledge bases, notes, documents, and meeting transcripts efficiently.
Run qmd search "my query" to search your indexed notes.
High-intent search queries and answers about installing and using this skill.
You can search local markdown notes using a hybrid approach combining BM25 keyword matching, vector search, and LLM reranking. This provides semantic understanding and precise retrieval for personal knowledge bases directly on your device.
Hybrid search combines keyword matching, semantic vector search, and LLM reranking to retrieve information. This mechanism ensures both exact term matching and conceptual understanding when finding notes and documents.
Yes, you need Node and SQLite to run this local search solution. These dependencies support the on-device indexing and retrieval processes required for querying markdown notes and meeting transcripts.
Yes, MCP integration allows you to search indexed markdown files from other applications. It supports both command-line interface and Multi-Channel Protocol connectivity for versatile knowledge retrieval workflows.
The best way to find specific concepts in large collections of meeting transcripts is using a local search engine with LLM reranking. It indexes text files and combines semantic understanding with keyword matching for comprehensive results.
peteromalletIndex and search local markdown and text documents with hybrid retrieval.
itsadijmbtSearch arXiv papers, download full text, and run local semantic search.
DrLuggelsDevelops RAG systems for LLM applications using vector databases and embedding models.
lanejSearch and audit personal knowledge bases with hybrid vector and BM25 indexing.
