Fetch MCP Server

Make HTTP requests and fetch web content.

Overview

Property	Value
Package	@modelcontextprotocol/server-fetch
Source	GitHub
Transport	stdio

Installation

npx -y @modelcontextprotocol/server-fetch

Configuration

mcp_servers:
  - name: fetch
    command: npx
    args: ["-y", "@modelcontextprotocol/server-fetch"]

With Custom User Agent

mcp_servers:
  - name: fetch
    command: npx
    args:
      - "-y"
      - "@modelcontextprotocol/server-fetch"
      - "--user-agent"
      - "AOF-Agent/1.0"

Tools

fetch

Fetch content from a URL.

Parameters:

• url (string, required): URL to fetch
• max_length (number, optional): Maximum content length to return
• start_index (number, optional): Start position for pagination
• raw (boolean, optional): Return raw HTML instead of markdown

Example:

{
  "tool": "fetch",
  "arguments": {
    "url": "https://api.example.com/status",
    "max_length": 5000
  }
}

Returns:

• For HTML pages: Converted markdown content
• For JSON APIs: Raw JSON response
• For other content: Raw text

Features

Automatic Content Conversion

The fetch server automatically:

• Converts HTML to readable markdown
• Preserves JSON structure
• Extracts text from common formats
• Handles character encoding

Pagination Support

For large content, use pagination:

{
  "tool": "fetch",
  "arguments": {
    "url": "https://example.com/long-page",
    "max_length": 5000,
    "start_index": 0
  }
}

Then fetch the next chunk:

{
  "tool": "fetch",
  "arguments": {
    "url": "https://example.com/long-page",
    "max_length": 5000,
    "start_index": 5000
  }
}

Raw HTML Mode

Get raw HTML when needed:

{
  "tool": "fetch",
  "arguments": {
    "url": "https://example.com/page",
    "raw": true
  }
}

Use Cases

API Monitor Agent

apiVersion: aof.sh/v1alpha1
kind: Agent
metadata:
  name: api-monitor
spec:
  model: google:gemini-2.5-flash
  mcp_servers:
    - name: fetch
      command: npx
      args: ["-y", "@modelcontextprotocol/server-fetch"]
  system_prompt: |
    You monitor API endpoints for health:
    1. Fetch status endpoints
    2. Check response codes
    3. Validate JSON structure
    4. Report anomalies

    Endpoints to monitor:
    - https://api.example.com/health
    - https://api.example.com/v1/status

Documentation Fetcher

apiVersion: aof.sh/v1alpha1
kind: Agent
metadata:
  name: doc-fetcher
spec:
  model: google:gemini-2.5-flash
  mcp_servers:
    - name: fetch
      command: npx
      args: ["-y", "@modelcontextprotocol/server-fetch"]
  system_prompt: |
    You fetch and summarize documentation:
    - Retrieve documentation pages
    - Extract key information
    - Summarize for users
    - Find relevant sections

Price Tracker

apiVersion: aof.sh/v1alpha1
kind: Agent
metadata:
  name: price-tracker
spec:
  model: google:gemini-2.5-flash
  mcp_servers:
    - name: fetch
      command: npx
      args: ["-y", "@modelcontextprotocol/server-fetch"]
  system_prompt: |
    You track prices from APIs:
    - Fetch pricing data
    - Compare historical prices
    - Alert on significant changes
    - Generate price reports

Webhook Tester

apiVersion: aof.sh/v1alpha1
kind: Agent
metadata:
  name: webhook-tester
spec:
  model: google:gemini-2.5-flash
  mcp_servers:
    - name: fetch
      command: npx
      args: ["-y", "@modelcontextprotocol/server-fetch"]
  system_prompt: |
    You test webhook endpoints:
    - Verify endpoint accessibility
    - Check response formats
    - Validate SSL certificates
    - Test authentication flows

Content Type Handling

Content Type	Handling
text/html	Converted to markdown
application/json	Returned as-is
text/plain	Returned as-is
text/markdown	Returned as-is
Other	Best-effort text extraction

Headers

The server sets default headers:

• User-Agent: Configurable via args
• Accept: text/html,application/json,text/plain,*/*

Limitations

• No POST/PUT/DELETE: Only GET requests
• No Custom Headers: Cannot set auth headers per-request
• No Cookies: Stateless requests only
• Size Limits: Large responses are truncated

Security

• URL Validation: URLs are validated before fetching
• No Local Files: Cannot fetch file:// URLs
• Timeout: Requests timeout after 30 seconds
• HTTPS Preferred: HTTP allowed but HTTPS recommended

Troubleshooting

Connection Refused

The target server may be blocking requests:

• Try with a different user agent
• Check if the site requires authentication
• Verify the URL is accessible

Timeout

Large pages may timeout:

• Use max_length to limit response size
• Try a more specific URL path
• Check network connectivity

Empty Response

Some sites return empty for bots:

• Site may require JavaScript rendering (use puppeteer instead)
• Site may block automated requests
• Check robots.txt compliance

SSL Errors

Certificate validation issues:

• Verify the site's SSL certificate is valid
• Check for certificate chain issues