Copy as Markdown

Turn any webpage into clean Markdown. Built for AI agents, pipelines, and anyone who needs structured content.

Why

When you feed web content to ChatGPT, Claude, Copilot, or any AI tool, raw HTML wastes tokens on noise — tags, styles, scripts, navigation. Markdown is what your AI actually reads.

Copy as Markdown fetches any URL and outputs clean, structured Markdown to stdout.

Usage

# No install required
npx copy-as-markdown https://example.com/article

# Save to file
npx copy-as-markdown https://example.com/article > article.md

# Pipe into an LLM
npx copy-as-markdown https://news.ycombinator.com/item?id=12345 | llm "summarize this"

# Include page title as H1
npx copy-as-markdown https://example.com/post --title

# Skip article extraction — convert the full page
npx copy-as-markdown https://example.com --no-readability

Options

| Flag | Description | | --- | --- | | --title | Prepend the page title as an H1 heading | | --no-readability | Skip Readability extraction — convert the full page body | | --help, -h | Show usage |

Install globally

npm install -g copy-as-markdown

How it works

1. Fetches the URL with a browser-like User-Agent
2. Extracts the main article content using @mozilla/readability
3. Converts the HTML to Markdown using Turndown.js
4. Outputs clean Markdown to stdout

Requires Node.js >= 18.

Browser Extension

Copy as Markdown also ships as a browser extension that converts selected web content in one keystroke (Cmd+Shift+M).

Works in Chrome, Edge, Brave, and Firefox.

👉 Get the extension on GitHub

Use cases

• Feed research and documentation into ChatGPT, Claude, or Copilot
• Build AI agent pipelines that process web content
• Populate RAG systems and vector databases with clean source material
• Automate content extraction in CI/CD or scripts
• Archive web pages as readable Markdown files

Privacy

• Zero telemetry — no analytics, no tracking
• Open source — read every line of code

License

MIT — github.com/sulmatajb/copy-as-markdown