@faircopy/markdown

Extract CommonMark and GitHub Flavored Markdown prose with locations in the original source. Supports .md and .markdown; MDX requires its own adapter.

npm i -D faircopy @faircopy/markdown

import { defineConfig } from '@faircopy/config'
import { markdown } from '@faircopy/markdown'

export default defineConfig({
  files: ['src/content/blog/**/*.md'],
  adapters: [markdown()],
  rules: { 'no-weasel-words': 'error' },
})

Headings, paragraphs, lists, quotes, link labels, emphasis, strikethrough, and table cells are copy. Inline markup stays within a prose segment, so a phrase split across emphasis or a link still reaches the rules as a sentence.

YAML/TOML frontmatter, fenced/indented/inline code, image syntax (including alt text), comments, link definitions/destinations/titles, and URL-looking text are excluded. Raw HTML blocks and inline HTML use @faircopy/html and its skip rules. Code and URLs create segment boundaries. Entities and escaped punctuation are decoded while source maps continue to point to the original Markdown. Character offsets use JavaScript UTF-16 code units, including CRLF and Unicode input.

Choose one representation per article for routine linting when HTML is generated from the Markdown. Both adapters report independently. When HTML is served and maintained independently, check both as separate surfaces and retain their file paths to distinguish findings. See the blog guide.