@kym6464/md-extract

Extract sections of a markdown file with a regular expression.

Installation

# Install globally
npm install -g @kym6464/md-extract

# Or use with npx (no installation required)
npx @kym6464/md-extract "pattern" document.md

Usage

Given a document called my-document.md:

# Welcome

This is my amazing markdown document.

## Extract me!

This section should be pulled out.

## Another section

This won't match the pattern.

You can extract the second section with the following command:

$ md-extract "Extract me!" my-document.md
## Extract me!

This section should be pulled out.

Options

Usage: md-extract [options] <pattern> <file>

Extract sections of a markdown file with a regular expression

Arguments:
  pattern                         Pattern to match against headings
  file                            Path to markdown file

Options:
  -V, --version                   output the version number
  -a, --all                       Print all matching sections (don't quit after first match)
  -s, --case-sensitive            Treat pattern as case sensitive
  -n, --no-print-matched-heading  Do not include the matched heading in the output
  --no-children                   Exclude content under child headings
  -h, --help                      display help for command

Examples

Extract all sections matching a pattern:

md-extract --all "^API" documentation.md

Case-sensitive matching:

md-extract --case-sensitive "TODO" project.md

Extract content without the heading:

md-extract --no-print-matched-heading "Summary" report.md

Extract only direct content, excluding child sections:

md-extract --no-children "Welcome" my-document.md

md-headings

List all headings in a markdown document, displayed as an indented tree.

$ md-headings my-document.md
Welcome
  Extract me!
  Another section

It also reads from stdin, so you can pipe content into it:

curl -sL https://example.com/doc.md | md-headings

Use --format json to get a nested tree with character counts per section (useful for AI agents deciding how to read content):

$ md-headings --format json my-document.md
[
  {
    "heading": "Welcome",
    "chars": 149,
    "ownChars": 49,
    "children": [
      { "heading": "Extract me!", "chars": 51 },
      { "heading": "Another section", "chars": 49 }
    ]
  }
]

Parent nodes include ownChars — the character count excluding children. Use this to predict how much content md-extract --no-children will return.

Or --format toon for a more compact representation:

$ md-headings --format toon my-document.md
[1]:
  - heading: Welcome
    chars: 149
    ownChars: 49
    children[2]{heading,chars}:
      Extract me!,51
      Another section,49

Options

Usage: md-headings [options] [file]

List headings in a markdown document

Arguments:
  file                  Path to markdown file (reads from stdin if omitted)

Options:
  -V, --version         output the version number
  -f, --format <type>   Output format: plain, json, toon (default: "plain")
  -h, --help            display help for command

Library Usage

You can also use this as a JavaScript library:

import { extractFromPath } from "@kym6464/md-extract";

const regex = /Extract me!/i;
const matches = await extractFromPath("my-document.md", regex);
console.log(matches); // Array of matching sections

Credits

This code was ported to JavaScript from sean0x42/markdown-extract.