qmd-skill-3（QMD 技能 3）

qmd - Quick Markdown Search

Local search engine for Markdown notes, docs, and knowledge bases. Index once, search fast.

When to use (trigger phrases)

• "search my notes / docs / knowledge base"
• "find related notes"
• "retrieve a markdown document from my collection"
• "search local markdown files"

Default behavior (important)

• Prefer qmd search (BM25). It's typically instant and should be the default.
• Use qmd vsearch only when keyword search fails and you need semantic similarity (can be very slow on a cold start).
• Avoid qmd query unless the user explicitly wants the highest quality hybrid results and can tolerate long runtimes/timeouts.

Prerequisites

• Bun >= 1.0.0
• macOS: brew install sqlite (SQLite extensions)
• Ensure PATH includes: $HOME/.bun/bin

Install Bun (macOS): brew install oven-sh/bun/bun

Install

bun install -g https://github.com/tobi/qmd

Setup

qmd collection add /path/to/notes --name notes --mask "**/*.md"
qmd context add qmd://notes "Description of this collection"  # optional
qmd embed  # one-time to enable vector + hybrid search

What it indexes

• Intended for Markdown collections (commonly **/*.md).
• In our testing, "messy" Markdown is fine: chunking is content-based (roughly a few hundred tokens per chunk), not strict heading/structure based.
• Not a replacement for code search; use code search tools for repositories/source trees.

Search modes

• qmd search (default): fast keyword match (BM25)
• qmd vsearch (last resort): semantic similarity (vector). Often slow due to local LLM work before the vector lookup.
• qmd query (generally skip): hybrid search + LLM reranking. Often slower than vsearch and may timeout.

Performance notes

• qmd search is typically instant.
• qmd vsearch can be ~1 minute on some machines because query expansion may load a local model (e.g., Qwen3-1.7B) into memory per run; the vector lookup itself is usually fast.
• qmd query adds LLM reranking on top of vsearch, so it can be even slower and less reliable for interactive use.
• If you need repeated semantic searches, consider keeping the process/model warm (e.g., a long-lived qmd/MCP server mode if available in your setup) rather than invoking a cold-start LLM each time.

Common commands

qmd search "query"             # default
qmd vsearch "query"
qmd query "query"
qmd search "query" -c notes     # Search specific collection
qmd search "query" -n 10        # More results
qmd search "query" --json       # JSON output
qmd search "query" --all --files --min-score 0.3

Useful options

• -n <num>: number of results
• -c, --collection <name>: restrict to a collection
• --all --min-score <num>: return all matches above a threshold
• --json / --files: agent-friendly output formats
• --full: return full document content

Retrieve

qmd get "path/to/file.md"       # Full document
qmd get "#docid"                # By ID from search results
qmd multi-get "journals/2025-05*.md"
qmd multi-get "doc1.md, doc2.md, #abc123" --json

Maintenance

qmd status                      # Index health
qmd update                      # Re-index changed files
qmd embed                       # Update embeddings

Keeping the index fresh

Automate indexing so results stay current as you add/edit notes.

• For keyword search (qmd search), qmd update is usually enough (fast).
• If you rely on semantic/hybrid search (vsearch/query), you may also want qmd embed, but it can be slow.

Example schedules (cron):

# Hourly incremental updates (keeps BM25 fresh):
0 * * * * export PATH="$HOME/.bun/bin:$PATH" && qmd update

# Optional: nightly embedding refresh (can be slow):
0 5 * * * export PATH="$HOME/.bun/bin:$PATH" && qmd embed

If your Clawdbot/agent environment supports a built-in scheduler, you can run the same commands there instead of system cron.

Models and cache

• Uses local GGUF models; first run auto-downloads them.
• Default cache: ~/.cache/qmd/models/ (override with XDG_CACHE_HOME).

Relationship to Clawdbot memory search

• qmd searches *your local files* (notes/docs) that you explicitly index into collections.
• Clawdbot's memory_search searches *agent memory* (saved facts/context from prior interactions).
• Use both: memory_search for "what did we decide/learn before?", qmd for "what's in my notes/docs on disk?".