Author here.

Quick taste - query your notes, pick the hits with fzf, bulk-tag them:

  fmql query ./vault 'type = "task" AND status = "open" AND priority > 2' \
    | fzf -m --preview 'head -40 {}' \
    | fmql append tags=triaged

Or follow references across files and get back deps graph you can render:

  fmql subgraph ./notes 'status = "active"' --follow blocked_by --format cytoscape

Origins: I run my own projects on workspaces of markdowns with yaml frontmatter (project board, documentation, configurations, agents definitions, ...). The format is great until you want to ask cross-file questions: what's stale, what's blocked on what, rename a field across 100+ files. I used to create one-off python scripts, until last week i decided to build a reusable tool that does all of it.

Beyond flat filters, there's a partial Cypher support for graph queries ("everything transitively blocking this sprint"), and you can swap in a semantic search backend via plugin when grep isn't enough. Core stays grep-only and dependency-light; semantic lives in a separate package.

A couple of things I'd be curious for feedback on:

- Linear scan is fine at a scale of few hundred documents, but if you have a larger vault and could try it, I'd love to hear when it breaks.

- What would you want to query your notes for that this doesn't handle? Genuinely interested in the use cases i haven't considered yet.

Repo: https://github.com/buyuk-dev/fmql Blog: https://www.buyuk.io/blog/fmql-launch/