Back to blog

StripFeed MCP Server: Give Your AI Coding Assistant Web Access

·3 min read·StripFeed Team

What is the StripFeed MCP Server?

The Model Context Protocol (MCP) lets AI coding assistants use external tools. The StripFeed MCP server gives your assistant the ability to read any web page as clean Markdown.

Instead of copying and pasting documentation into your chat, just ask your assistant to fetch it:

"Read the React useEffect documentation and explain the cleanup function"

Your assistant calls StripFeed, gets clean Markdown, and answers your question with up-to-date information from the actual docs.

Setup

Claude Code

One command:

claude mcp add stripfeed -- npx -y @stripfeed/mcp-server

Set your API key:

export STRIPFEED_API_KEY=sf_live_your_key

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "stripfeed": {
      "command": "npx",
      "args": ["-y", "@stripfeed/mcp-server"],
      "env": {
        "STRIPFEED_API_KEY": "sf_live_your_key"
      }
    }
  }
}

Windsurf

Add to your MCP configuration:

{
  "mcpServers": {
    "stripfeed": {
      "command": "npx",
      "args": ["-y", "@stripfeed/mcp-server"],
      "env": {
        "STRIPFEED_API_KEY": "sf_live_your_key"
      }
    }
  }
}

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "stripfeed": {
      "command": "npx",
      "args": ["-y", "@stripfeed/mcp-server"],
      "env": {
        "STRIPFEED_API_KEY": "sf_live_your_key"
      }
    }
  }
}

Available Tools

The MCP server exposes three tools:

fetch_url

Fetch a single URL and convert to Markdown.

Parameters:

  • url (required) - The URL to fetch
  • format - Output format: markdown, json, text, html
  • selector - CSS selector to extract specific content
  • model - AI model ID for cost tracking
  • cache - Enable/disable caching (default: true)
  • ttl - Cache TTL in seconds
  • max_tokens - Maximum tokens in output

batch_fetch

Fetch up to 10 URLs in parallel (Pro plan).

Parameters:

  • urls (required) - Array of URLs to fetch
  • model - AI model ID for cost tracking

check_usage

Check your current month's usage, plan limits, and remaining requests.

Usage Examples

Once installed, just ask your assistant naturally:

Read documentation:

"Fetch the Next.js App Router docs and explain how layouts work"

Compare libraries:

"Fetch the Prisma and Drizzle ORM documentation and compare their query APIs"

Research before coding:

"Read https://stripe.com/docs/api/charges and help me implement a payment flow"

Check multiple sources:

"Fetch React, Vue, and Svelte docs and compare their state management approaches"

Extract specific content:

"Fetch the pricing table from https://example.com/pricing using the selector .pricing-grid"

Response Details

Every fetch includes metadata your assistant can use:

  • Token count - How many tokens the clean content uses
  • Savings percentage - How much smaller Markdown is vs raw HTML
  • Cache status - Whether the result came from cache
  • Fetch time - How long the request took in milliseconds

Get Started

  1. Sign up for free to get your API key
  2. Run the setup command for your tool
  3. Start asking your assistant to read web pages

The free plan includes 200 requests/month. That's enough for daily use with your coding assistant.

npm: @stripfeed/mcp-server | GitHub: StripFeed/mcp-server