Search local knowledge bases with hybrid keyword, vector, and LLM reranking.
npx skills add https://github.com/kwasi-cpu/hermes-agent --skill qmd-kwasi-cpu
Please help me install this Agent Skill. Skill: qmd Source: https://github.com/kwasi-cpu/hermes-agent/tree/main/optional-skills/research/qmd Command: npx skills add https://github.com/kwasi-cpu/hermes-agent --skill qmd-kwasi-cpu
This Skill provides a powerful, local AI-powered search engine for your personal knowledge base, notes, and documents, eliminating the need for cloud dependencies.
Use the qmd skill to search your notes for information about the Q3 project roadmap.
High-intent search queries and answers about installing and using this skill.
Local semantic search processes your personal notes entirely on-device by combining vector embeddings with keyword matching, eliminating cloud dependencies while keeping data private and secure.
Natural language queries retrieve information from markdown knowledge bases using a hybrid retrieval engine that applies vector search, keyword matching, and LLM reranking for highly accurate results.
Local hybrid search supports seamless agent interaction by integrating directly with command line interface (CLI) and Model Context Protocol (MCP) environments for automated knowledge base querying.
Searching thousands of meeting transcripts and project documentation locally is best handled by a hybrid retrieval engine combining BM25 keyword matching, vector search, and LLM reranking.
On-device knowledge base search requires no cloud dependencies or external APIs, as all models and data are stored and processed locally on your machine for complete offline operation.
LLM reranking improves local knowledge base search accuracy by reordering initial vector and keyword search results, ensuring the most contextually relevant documents surface for complex queries.
nelohenriqIndex local markdown notes and transcripts for hybrid on-device search via MCP.
thecloudtipsSynthesize sourced answers with inline citations and confidence ratings.
ivanshamaevBuild RAG ingestion and query pipelines with hybrid retrieval and reranking.
shinnytechBuild RAG systems with vector databases and semantic search.
