Documentation

Installation

Use curl or any HTTP client to fetch Markdown content:

Terminal

curl https://api.mdfetch.ai/https://github.com/user/repo \
-H "Authorization: Bearer YOUR_TOKEN"

Install MdFetch as an MCP server in Claude Code:

Terminal

claude mcp add --transport http mdfetch https://mcp.mdfetch.ai \
--header "Authorization: Bearer YOUR_TOKEN"

Ask Claude to fetch any page, or force it with mdfetch:

Example

mdfetch https://shopify.dev/docs/storefronts/themes to help me build a Shopify theme

💡 Pro tip

Add this to your Claude Code instructions for automatic fetching:

Instructions

- ALWAYS call the mdfetch tool to fetch webpages in Markdown format.

Add this to your MCP settings file:

mcp_settings.json

{
  "mcpServers": {
    "mdfetch": {
      "url": "https://mcp.mdfetch.ai",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Add this to your MCP settings file:

mcp_settings.json

{
  "mcpServers": {
    "mdfetch": {
      "url": "https://mcp.mdfetch.ai",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

Examples

Fetch Documentation

Terminal

curl https://api.mdfetch.ai/https://hexdocs.pm/phoenix/overview.html \
-H "Authorization: Bearer YOUR_TOKEN"

Fetch GitHub Repository

MdFetch automatically fetches the README and repository structure:

Terminal

curl https://api.mdfetch.ai/https://github.com/phoenixframework/phoenix \
-H "Authorization: Bearer YOUR_TOKEN"

Fetch Source Code Files

Get raw source code from GitHub, GitLab, or any supported platform:

Terminal

curl https://api.mdfetch.ai/https://github.com/elixir-lang/elixir/blob/main/lib/elixir/lib/enum.ex \
-H "Authorization: Bearer YOUR_TOKEN"

API Reference

Endpoint

GET https://api.mdfetch.ai/{url}

Replace {url} with the full URL you want to fetch.

Authentication

Include your API token in the Authorization header:

Header

Authorization: Bearer YOUR_TOKEN

Rate Limits

Monitor usage in your dashboard
See the homepage for pricing details

Volume Pricing

For high-volume production workloads, we offer custom volume pricing with:

Volume discounts competitive with other providers
Dedicated support and SLA guarantees available
Custom rate limits tailored to your needs
Flexible billing options (monthly, quarterly, annual)

Response Format

Returns clean Markdown as plain text with UTF-8 encoding:

Response

Content-Type: text/markdown; charset=utf-8

Error Codes

401 - Invalid or missing authentication token
429 - Rate limit exceeded
500 - Server error or failed to fetch URL

Best Practices

🔒 Token Security

Never commit tokens to version control
Use environment variables for token storage
Regenerate tokens if compromised (contact support)
Keep tokens private and don't share them

⚡ Performance Tips

Cache responses when possible to save requests
Monitor your usage in the dashboard
Use specific URLs rather than entire websites
MdFetch automatically optimizes content size

🎯 Content Optimization

GitHub/GitLab files are fetched as raw source
Documentation sites preserve structure and code blocks
HTML is stripped, keeping only content
Results are optimized for LLM context windows

Frequently Asked Questions

What sites are supported?

MdFetch works with any publicly accessible webpage. It has special optimizations for GitHub, GitLab, documentation sites, and platforms that serve raw content.

Can I use it commercially?

Yes! MdFetch is suitable for both personal and commercial use. Just keep your API token secure.

How much smaller is the output?

Typically 70-90% smaller than raw HTML. This means significantly fewer tokens when sending to LLMs, saving you money and fitting more content in context windows.

What if I need help?

Ready to get started?

Get Started Free