Web2MD — open source

Two packages live here, both MIT:

Both are extracted from Web2MD, a Chrome extension that converts web pages to Markdown for AI tools.

web2md-mcp-server

An MCP server exposing 12 tools to any MCP client — Claude Code, Claude Desktop, Cursor, Cline, and others.

npx -p web2md-mcp-server web2md-mcp-setup

That writes the config for your client and, if a tool needs an account, opens the browser for OAuth — no password is typed anywhere.

Or configure it by hand (the binary is web2md-mcp):

{
  "mcpServers": {
    "web2md": { "command": "npx", "args": ["-y", "-p", "web2md-mcp-server", "web2md-mcp"] }
  }
}

Tools: convert_url, convert_file, semantic_search, get_conversion, bridge_convert_url, agent_convert, agent_batch_convert, list_skills, get_skill, search_library, get_library_items, sync_library.

Two of these are worth calling out:

• bridge_convert_url connects to your own signed-in Chrome session, so an agent can read pages that need a login or run heavy JavaScript — the ones a server-side fetch gets a login wall or an empty shell for.
• agent_batch_convert takes up to 50 URLs in one call and returns Markdown plus token counts for each.

convert_url works without an account. Tools that touch your library or the hosted API need a free key — the setup command above handles that, or set WEB2MD_API_KEY yourself.

web2md-core

Turn messy HTML into clean, LLM-ready Markdown.

This is the extraction and conversion engine behind Web2MD. It is a pure library — give it an HTML string, get Markdown back. No network calls, no API key, no account. Nothing in this package talks to a server.

npm install web2md-core

Why convert at all

Feeding raw HTML to a language model wastes most of your context window on markup, navigation, and ads. Converting first cuts that down and gives the model a document it can actually follow.

The library reports both numbers so you can see the difference:

import { convertToMarkdown } from 'web2md-core'

const result = convertToMarkdown(html, { url: 'https://example.com/post' })

console.log(result.markdown)
console.log(result.metadata.originalTokenCount, '→', result.metadata.tokenCount)
// e.g. 417 → 281

convertToMarkdown returns null when it cannot find a main content block — check for that rather than assuming a result.

What it does

• Finds the actual article. Strips navigation, sidebars, ads, cookie banners, and footers, keeping the content a reader came for.
• Preserves structure. Headings, lists, tables, and fenced code blocks survive the round trip — that structure is what lets a model answer questions about one specific section.
• Reports tokens. Estimated counts for both the original HTML and the cleaned Markdown, plus helpers to split or trim for a target context window.
• Runs anywhere. Uses linkedom for parsing, so it works in Node without a browser.

API

convertToMarkdown(html, options?)

The main entry point. Note the signature takes two arguments — the URL goes inside options, not as a positional parameter:

convertToMarkdown(html, { url: 'https://example.com/post' })

... View full README on GitHub