Skip to content

[GitHub Trending] tobi/qmd

8.9 relevance
Score Breakdown
technical depth
7
novelty
8
actionability
9
community
8
strategic
7
personal
9

Scored daily by a customisable AI persona to surface the most relevant engineering leadership news.

A local-first CLI search engine for personal knowledge bases, aligning with productivity tooling and privacy-focused trends.

2026-04-07 DevTools github.com
mini cli search engine for your docs, knowledge bases, meeting notes, whatever. Tracking current sota approaches while being all local - tobi/qmd
Summary

QMD is an on-device search engine that indexes markdown and documents using a hybrid stack of BM25, vector semantic search, and LLM reranking via node-llama-cpp with GGUF models. It exposes an MCP server with tools like query, get, and multi_get, supporting HTTP transport for persistent servers and structured JSON/file outputs optimized for agentic workflows.

Key Takeaways
  • Integrate QMD's MCP server into your agent workflows to enable hybrid search over local knowledge bases with GGUF model reranking, avoiding external API calls.
Why it matters

As a senior engineer building AI agent orchestration systems, you need a local, standards-based knowledge retrieval layer that integrates via MCP without cloud dependencies, which QMD provides through its hybrid search and on-device LLM reranking.