qmd
Hacker ApprovedMarkdown search tool in Developer tools
Local Markdown search mixing keyword, vector and LLM reranking, popular as agent memory. Caveat: designed around single-machine setups and downloads several large local models.
Local Markdown search mixing keyword, vector and LLM reranking, popular as agent memory. Caveat: designed around single-machine setups and downloads several large local models.
20 people gave an opinion
- recommend it
- 19
- criticize it
- 1
- approval share
- 95%
Recommended in 9 of 13 months
- November 2025: not recommended
- December 2025: not recommended
- January 2026: recommended
- February 2026: recommended
- March 2026: recommended
- April 2026: recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: not recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it carries the badge
- 19 people recommend itNeeds 10 or more, met.
- 19 recommenders for every criticNeeds 2 to 1 or better, met.
- Recommended in 9 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Good, hybrid search results
“My agents use both depending on what type of search they are doing, and both provide good results.”
Comment from 9 Jun 2026 in “Is Grep All You Need? How Agent Harnesses Reshape Agentic Search” “The hybrid BM25 + vector + LLM reranking approach is exactly what I've been wanting for my notes.”
Comment from 26 Jan 2026 in “QMD - Quick Markdown Search” “you still get semantic search to help you find the right documents quickly”
Comment from 27 Oct 2026 in “Survival LLM: AI assistant and emergency library that runs on your own computer” - Comment from 15 Jan 2026 in “Ask HN: How are you doing RAG locally?”
Runs locally with local models
“tools like QMD that beautifully download and use local models under the hood”
Comment from 11 May 2026 in “Local AI needs to be the norm” “I tried to make it not overkill and keep things local.”
Comment from 26 Jan 2026 in “QMD - Quick Markdown Search” Default RAG engine for OpenClaw
“now it's used as the defacto RAG engine for OpenClaw”
Comment from 13 Mar 2026 in “Launch HN: Captain (YC W26) – Automated RAG for Files” “it's a great option for OpenClaw!”
Comment from 13 Mar 2026 in “Launch HN: Captain (YC W26) – Automated RAG for Files” Clever chunking and indexing approach
“It uses some cool techniques like resilient chunking and embedding”
Comment from 17 Feb 2026 in “Sub-Millisecond RAG on Apple Silicon. No Server. No API. One File” “The QMD indexing approach is clever”
Comment from 14 Feb 2026 in “Show HN: Ghost – Session memory for Claude Code (local, qmd, Git-integrated)”
What people warn about
Each mark is one comment
Built around single-machine setups
“qmd is focussed on local to the point of designing around single machine setups”
Comment from 15 Jul 2026 in “Show HN: Capn-hook for coding agents – don't grep the same mystery twice” Requires several large embedding models
“though it requires installing 3 local embedding models (300MB, 640MB, 1.1GB)”
Comment from 27 Mar 2026 in “From zero to a RAG system: successes and failures”
What people said
17 of the 21 comments that mention it, newest first. Each links to the original comment.
“you still get semantic search to help you find the right documents quickly”
“QMD is great! Concept is similar, relies on having an existing Markdown store for your agent's memory.”
“qmd is focussed on local to the point of designing around single machine setups”
“one of the reasons I built gmd to replace qmd”
“My agents use both depending on what type of search they are doing, and both provide good results.”
“tools like QMD that beautifully download and use local models under the hood”
“Run QMD on an obsidian vault and that's like 80% of the way there”
“I ended up using QMD + custom skills to keep it up to date as new data comes in.”
“after the third injected wrong fact I went back to QMD and prose / summarization”
“though it requires installing 3 local embedding models (300MB, 640MB, 1.1GB)”
“it's a great option for OpenClaw!”
“now it's used as the defacto RAG engine for OpenClaw”
“It uses some cool techniques like resilient chunking and embedding”
“The QMD indexing approach is clever”
“The hybrid BM25 + vector + LLM reranking approach is exactly what I've been wanting for my notes.”
“I tried to make it not overkill and keep things local.”
“this tool gives better results and enables even more fuzzy queries”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.