Phrasing

Phrasing is a lightweight static app for browsing and searching phrase collections stored as Markdown. Built-in corpora ship with the site; you can switch datasets in the browser or upload your own .md files without a page reload.

What it does

The build pipeline converts Markdown under data/builtin/ into a searchable static site. Fuse.js handles fuzzy search; a fixed sidebar provides category navigation and A–Z section links. The same parsing logic runs at build time and in the browser via shared content-core.js, so uploaded files behave identically to built-in datasets.

Input format stays deliberately simple:

# Category name

## A

- First phrase
- Second phrase

A single file can contain multiple # sections. List items under ## letter headings become searchable entries.

Built-in datasets

Dataset	Contents
phrasing	~15,000 entries from Fifteen Thousand Useful Phrases (Grenville Kleiser, 1910)
insults	Shakespeare insults, grouped by type

Add a folder under data/builtin/ and rebuild — the manifest picks it up automatically.

Why “Phrasing”

The default corpus comes from a Project Gutenberg find: a 1910 reference book of business expressions, conversational phrases, impressive formulations, and more. Under Conversational — I, you might find:

I am anxious to discharge the very onerous debt I owe you.

Most entries are archaic or slightly questionable by modern standards, which is part of the charm — a searchable snapshot of how people were told to sound articulate a century ago.

Current features

• Dataset switching — toggle between built-in corpora or upload custom Markdown via the data control
• Instant search — Fuse.js fuzzy matching across all indexed phrases
• Themes — Dracula, Cursor, Orangde, Black & White, Light, Sepia, and Rose
• Static hosting — npm run build outputs to dist/; deploys to GitHub Pages on push via Actions

Stack: Node.js build script, markdown-it, Fuse.js, optional Express dev server.