qmd

Index and search local Markdown notes with hybrid retrieval.

Updated Aug 27, 2026

Search Tags:#MCP #retrieval #BM25 #knowledge-base #vector-search #local-search #LLM-reranking

Ready to run

npx skills add https://github.com/wwwillott/jobnimbus --skill qmd-wwwillott

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: qmd
Source: https://github.com/wwwillott/jobnimbus/tree/main/optional-skills/research/qmd
Command: npx skills add https://github.com/wwwillott/jobnimbus --skill qmd-wwwillott

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Locally index and search Markdown notes, documents, and transcripts without cloud dependencies.

Core Features & Use Cases

Hybrid search combines BM25 keyword search, vector similarity, and LLM reranking for relevant results.
On-device indexing builds fast, private indexes of local collections like notes, docs, and transcripts.
CLI and MCP integration supports command-line queries and HermesAgent MCP workflows.

Quick Start

Index a local folder with qmd and run a search across your collections.

Frequently Asked Questions about qmd

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I search Markdown notes and documents locally without cloud dependencies?▼

Local on-device search solves this by indexing Markdown notes and documents directly on your machine. It uses hybrid retrieval combining BM25 keyword search, vector similarity, and LLM reranking, operating entirely offline without any cloud dependencies.

What is hybrid retrieval for personal knowledge bases?▼

Hybrid retrieval combines BM25 keyword search, vector similarity, and LLM reranking to provide relevant results from personal knowledge bases. This mechanism ensures highly accurate search outcomes across locally indexed notes, documents, and transcripts.

Can I run offline vector search across private transcripts on my machine?▼

Yes, you can run offline vector search across private transcripts on your machine. The system provides on-device indexing with vector similarity and operates entirely offline, ensuring your private transcripts remain secure without any cloud connections.

How do I integrate local search with an MCP workflow?▼

You integrate local search with an MCP workflow using the provided MCP interface. This supports HermesAgent MCP workflows alongside a CLI, allowing you to run queries across your locally indexed collections seamlessly within your agent architecture.

Does on-device BM25 search work for large local document collections?▼

On-device BM25 search works for large local document collections by building fast, private indexes of local files. It is designed for developers, researchers, and knowledge workers needing efficient offline retrieval across extensive personal knowledge bases.