mreader

Extract clean Markdown from any URL. Prefers server-side Markdown via content negotiation (Accept: text/markdown, e.g. Cloudflare Docs as Markdown), falls back to client-side extraction with @extractus/article-extractor and Turndown.

Install

npm install -g mreader

CLI

# Print article as Markdown (with YAML frontmatter)
mreader https://example.com/article

# Save to file
mreader https://example.com/article -o article.md

# Output as JSON (title, author, content, excerpt, …)
mreader https://example.com/article -j

# Skip YAML frontmatter
mreader https://example.com/article --no-frontmatter

# Skip content negotiation, always extract from HTML
mreader https://example.com/article --no-content-negotiation

# Read HTML from stdin
curl -s https://example.com/article | mreader - --url https://example.com/article

Options

| Flag | Description | |------|-------------| | -o, --output <file> | Write output to file instead of stdout | | -j, --json | Output as JSON | | --no-frontmatter | Omit YAML frontmatter from output | | --no-content-negotiation | Skip Accept: text/markdown, always extract from HTML | | --url <url> | Base URL for stdin mode | | -h, --help | Show help | | -v, --version | Show version |

Content negotiation

By default mreader sends Accept: text/markdown first. If the server responds with Content-Type: text/markdown, that response is used directly — no extraction or conversion needed. Otherwise mreader falls back to fetching the HTML and extracting the article client-side.

Use --no-content-negotiation (CLI) or noContentNegotiation: true (API) to skip this step.

API

import { read, readFromHtml, formatArticle } from "mreader";

// From URL
const article = await read("https://example.com/article");
console.log(article.title);
console.log(article.content); // clean Markdown

// From HTML string
const article2 = await readFromHtml(html, "https://example.com/article");

// Format with YAML frontmatter
console.log(formatArticle(article));

// Format without frontmatter
console.log(formatArticle(article, false));

read(url, options?)

Fetches the URL and returns an Article object.

Options:

interface ReadOptions {
  headers?: Record<string, string>;
  signal?: AbortSignal;
  noContentNegotiation?: boolean; // skip Accept: text/markdown
}

readFromHtml(html, url)

Parses an HTML string and returns an Article object.

formatArticle(article, frontmatter?)

Returns the article as a Markdown string, optionally prefixed with YAML frontmatter (default: true).

Article

interface Article {
  title: string;
  author: string | null;
  content: string; // Markdown
  excerpt: string | null;
  published: string | null;
  source: string | null;
  url: string;
}

License

MIT

Made with ❤️ by the Roman Ožana