Web MCP Server

A Model Context Protocol (MCP) server for web search and URL content retrieval, built with FastAPI and fastapi-mcp.

Features

• Web Search: Search the web using DuckDuckGo
• URL Content Retrieval: Fetch and parse content from any URL as markdown
  • HTML pages: Convert to clean markdown format
  • Files: Extract text from various file types via Goggles API (PDF, DOCX, etc.)

Running Locally

Prerequisites

• Python 3.13+
• Virtual environment (recommended)

Setup

# Create and activate virtual environment
python -m venv .venv
.venv\Scripts\Activate.ps1  # Windows PowerShell
# or
source .venv/bin/activate    # Linux/Mac

# Install dependencies
pip install -r requirements.txt

# Run the server
python app.py

The server will be available at:

• API: http://localhost:8000
• API Docs: http://localhost:8000/docs
• MCP Endpoint: http://localhost:8000/mcp

Running with Docker

Build and run with Docker Compose (Recommended)

# Build and start the container
docker-compose up -d

# View logs
docker-compose logs -f

# Stop the container
docker-compose down

Build and run with Docker directly

# Build the image
docker build -t web-mcp-server .

# Run the container
docker run -d -p 8000:8000 --name web-mcp-server web-mcp-server

# View logs
docker logs -f web-mcp-server

# Stop the container
docker stop web-mcp-server
docker rm web-mcp-server

API Endpoints

/search - Web Search

Search the web using DuckDuckGo.

Parameters:

• query (string, required): The search query

Example:

curl "http://localhost:8000/search?query=python+programming"

/content - Get URL Content

Retrieve and parse content from a URL, returned as markdown with pagination support.

Parameters:

• url (string, required): The URL to fetch content from
• offset (integer, optional): Character offset to start from (default: 0)
• limit (integer, optional): Maximum characters to return (default: 10000, max: 50000)

Examples:

# Get first 10,000 characters (default)
curl "http://localhost:8000/content?url=https://example.com"

# Get specific range
curl "http://localhost:8000/content?url=https://example.com&offset=10000&limit=10000"

# Get more content (up to 50,000 characters)
curl "http://localhost:8000/content?url=https://example.com&limit=50000"

Response includes pagination info:

• total: Total content length
• returned: Characters returned in this response
• has_more: Whether more content is available

MCP Integration

Connect this server to any MCP client (Claude Desktop, Cursor, Windsurf) using:

{
  "mcpServers": {
    "web-mcp": {
      "url": "http://localhost:8000/mcp"
    }
  }
}

Configuration

Goggles API Endpoint

The Goggles API endpoint can be configured via the GOGGLES_URL environment variable.

Default: http://localhost:8001

Set via environment variable:

# Linux/Mac
export GOGGLES_URL=http://your-goggles-instance:8001

# Windows PowerShell
$env:GOGGLES_URL="http://your-goggles-instance:8001"

# Docker
docker run -e GOGGLES_URL=http://your-goggles-instance:8001 -p 8000:8000 web-mcp-server

# Docker Compose
# Create a .env file in the project root:
echo "GOGGLES_URL=http://your-goggles-instance:8001" > .env

Docker Hub Deployment

This repository includes a GitHub Actions workflow that automatically builds and pushes Docker images to Docker Hub when code is pushed to the main branch.

Setup GitHub Secrets and Variables

To enable automatic Docker Hub deployment, configure the following in your GitHub repository settings:

Secrets (Settings → Secrets and variables → Actions → New repository secret):

• DOCKER_PASSWORD: Your Docker Hub password or access token

Variables (Settings → Secrets and variables → Actions → New repository variable):

• DOCKER_USERNAME: Your Docker Hub username

Once configured, every push to main will automatically build and push a new Docker image tagged with both latest and the commit SHA.

Dependencies

• fastapi - Web framework
• fastapi-mcp - MCP integration
• uvicorn - ASGI server
• requests - HTTP client
• beautifulsoup4 - HTML parsing
• markdownify - HTML to markdown conversion