markd-mcp
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@markd-mcpextract clean markdown from https://example.com/article"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Markd MCP Server
An official Model Context Protocol (MCP) server that brings Markd straight into your AI coding workflow.
🚀 Why this exists: stop burning context tokens
Ask an AI like Claude or GPT-4 to read a webpage, and it usually ends up chewing through raw HTML. That raw HTML can easily run 25,000 to 30,000 tokens once you count inline styles, tracking scripts, and bloated DOM markup.
Markd fixes that. It spins up a headless Chrome browser (via Puppeteer) in the background, fully renders the page (JavaScript included), and pulls out just the real content using Mozilla Readability and Turndown.
That same article, converted to clean Markdown, comes out to around 2,000 to 2,500 tokens. That's roughly a 90% drop in token usage every time your AI pulls something off the web. Faster responses, cheaper API bills, and a lot less noise clogging up the context window.
Related MCP server: Skrape MCP Server
🛠️ What it gives your AI
One tool, does the job well:
extract_markdownInput:
{ url: "string" }What it does: fetches the URL, renders it fully, and hands back the main article as clean Markdown.
🌍 How to Use (NPM Package)
The fastest and most reliable way to use Markd-MCP across any machine (Windows, Mac, Linux) is to install it globally. This guarantees the headless Chrome browser is installed properly and makes the tool instant for your AI.
1. Install Globally
Open your terminal and run:
npm install -g markd-mcp2. Add to Claude Code
Add this to your claude.json (or use claude mcp add):
{
"mcpServers": {
"markd-mcp": {
"command": "markd-mcp",
"args": []
}
}
}💻 Running it locally (For Developers)
If you want to modify the code or run it from source.
1. Install
git clone https://github.com/mayankfulwariya/markd-mcp.git
cd markd-mcp
npm install
npx puppeteer browsers install chrome2. Point your AI client at it
Claude Code
Add this to your claude.json, or use the claude mcp add command:
{
"mcpServers": {
"markd-mcp": {
"command": "node",
"args": ["/absolute/path/to/markd-mcp/src/index.js"]
}
}
}Claude Desktop
Same snippet, different file:
Mac:
~/Library/Application Support/Claude/claude_desktop_config.jsonWindows:
%APPDATA%\Claude\claude_desktop_config.json
Just swap in the real path to wherever you cloned the repo.
Under the hood
Puppeteer – renders JS-heavy sites (React, Next.js docs, that kind of thing) fully before anything gets extracted.
Mozilla Readability – the same engine behind Firefox's Reader View. Strips out sidebars, headers, footers, ads, all the junk.
Turndown – turns the clean HTML into proper Markdown.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/mayankfulwariya/markd-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server