Retrieve relevant markdown notes and transcripts using hybrid keyword and semantic search.
npx skills add https://github.com/ruiyangruiyi/hermes-agent --skill qmd-ruiyangruiyi
Please help me install this Agent Skill. Skill: qmd Source: https://github.com/ruiyangruiyi/hermes-agent/tree/main/optional-skills/research/qmd Command: npx skills add https://github.com/ruiyangruiyi/hermes-agent --skill qmd-ruiyangruiyi
qmd eliminates the frustration of finding the right information across large personal collections of markdown notes, documentation, and meeting transcripts by providing fast keyword search and higher-quality semantic retrieval locally.
Use Case Example: You have hundreds of meeting transcripts and project notes; you can ask for the decisions made about the database migration and retrieve the most relevant excerpts even if the wording varies across documents.
Use Hermes to query your local knowledge base by running: qmd query "what decisions were made about the database migration".
High-intent search queries and answers about installing and using this skill.
Hybrid keyword and semantic retrieval indexes your local markdown notes and transcripts on-device, using BM25 exact-match combined with vector search and LLM reranking to surface relevant documents without cloud dependencies.
Hybrid RAG combines BM25 keyword search for exact terms with vector semantic search for conceptual queries, then applies LLM reranking. This dual approach improves retrieval quality for ambiguous queries across personal markdown collections.
Yes, you can index and query your personal knowledge bases entirely on-device. Local embedding generation and retrieval make hybrid search suitable for private or offline workflows without requiring cloud dependencies.
Yes, Node.js is required on the host to run local collection indexing and embedding generation. On macOS, optional SQLite extension support is available for enhanced functionality.
You can query your local notes by running the qmd CLI command directly in the terminal, or by exposing searchable tools to Hermes Agent through an MCP server interface for integration.
Local AI semantic search requires on-device processing for embedding generation, meaning performance depends on your hardware. It also requires Node.js and optional SQLite extension support on macOS for full functionality.
cxnaiveIndex local markdown notes and perform hybrid semantic search with BM25 and vector embeddings.
cdubiel08Index markdown documents into a SQLite database with full-text search and semantic embeddings.
jacardlIndex and search local knowledge bases with hybrid retrieval.
evolv3aiCreate RAG systems with Google Gemini's managed File Search API.
