FetchV2 MCP Server

Model Context Protocol (MCP) server for web content fetching and extraction.

This MCP server provides tools to fetch webpages, extract clean content using Trafilatura, and discover links for batch processing.

Features

• Fetch Webpages: Extract clean markdown content from any URL
• Batch Fetching: Fetch up to 10 URLs in a single request
• Link Discovery: Find and filter links on any webpage
• llms.txt Support: Parse and fetch LLM-friendly documentation indexes
• Smart Extraction: Trafilatura removes boilerplate (navbars, ads, footers)
• Robots.txt Compliance: Respects robots.txt with graceful timeout handling
• Pagination Support: Handle large pages with start_index parameter

Prerequisites

1. Install uv from Astral
2. Install Python 3.10 or newer using uv python install 3.10

Installation

Cursor	VS Code
Install MCP Server	Install on VS Code

Or configure manually in your MCP client:

{
  "mcpServers": {
    "fetchv2": {
      "command": "uvx",
      "args": ["fetchv2-mcp-server@latest"],
      "disabled": false,
      "autoApprove": []
    }
  }
}

Config file locations:

• Claude Desktop (macOS): ~/Library/Application Support/Claude/claude_desktop_config.json
• Claude Desktop (Windows): %APPDATA%\Claude\claude_desktop_config.json
• Windsurf: ~/.codeium/windsurf/mcp_config.json
• Kiro: .kiro/settings/mcp.json in your project

Install from PyPI

# Using uv
uv add fetchv2-mcp-server

# Using pip
pip install fetchv2-mcp-server

Basic Usage

Example prompts to try:

• "Fetch the documentation from <URL>"
• "Find all links on <docs URL> that contain 'tutorial'"
• "Read these three pages and summarize the differences: [url1, url2, url3]"

Available Tools

fetch

Fetches a webpage and extracts its main content as clean markdown.

fetch(url: str, max_length: int = 5000, start_index: int = 0) -> str

Parameter	Type	Default	Description
url	str	required	The webpage URL to fetch
max_length	int	5000	Maximum characters to return
start_index	int	0	Character offset for pagination
get_raw_html	bool	false	Skip extraction, return raw HTML
include_metadata	bool	true	Include title, author, date
include_tables	bool	true	Preserve tables in markdown
include_links	bool	false	Preserve hyperlinks
bypass_robots_txt	bool	false	Skip robots.txt check

fetch_batch

Fetches multiple webpages in a single request.

fetch_batch(urls: list[str], max_length_per_url: int = 2000) -> str

Parameter	Type	Default	Description
urls	list[str]	required	List of URLs (max 10)
max_length_per_url	int	2000	Character limit per URL
get_raw_html	bool	false	Skip extraction for all URLs

discover_links

Discovers all links on a webpage with optional filtering.

discover_links(url: str, filter_pattern: str = "") -> str

Parameter	Type	Default	Description
url	str	required	The webpage URL to scan
filter_pattern	str	""	Regex to filter links (e.g., /docs/)

fetch_llms_txt

Fetch and parse an llms.txt file to discover LLM-friendly documentation.

fetch_llms_txt(url: str, include_content: bool = False) -> str

Parameter	Type	Default	Description
url	str	required	URL to an llms.txt file
include_content	bool	false	Also fetch content of all linked pages
max_length_per_url	int	2000	When include_content=True, max chars per page

⚠️ Important: By default, only the llms.txt index is fetched — the linked markdown files are NOT downloaded to context. Set include_content=True to explicitly fetch all linked pages.

Example:

# DEFAULT: Only fetches the index (lightweight, ~1KB)
fetch_llms_txt(url="https://docs.example.com/llms.txt")
# Returns: title + list of links with descriptions

# EXPLICIT: Fetches index + all linked .md files (can be large)
fetch_llms_txt(url="https://docs.example.com/llms.txt", include_content=True)
# Returns: structure + content of all linked pages

Note: Relative URLs (e.g., /docs/guide.md) are automatically resolved to absolute URLs.

Workflow Example

Step 1: Discover relevant documentation pages

discover_links(url="https://docs.example.com/", filter_pattern="/guide/")

Step 2: Batch fetch the pages you need

fetch_batch(urls=["https://docs.example.com/guide/intro", "https://docs.example.com/guide/setup"])

Prompts

• fetch_manual - User-initiated fetch that bypasses robots.txt
• research_topic - Research a topic by fetching multiple relevant URLs

Development

# Clone and install
git clone https://github.com/praveenc/fetchv2-mcp-server.git
cd fetchv2-mcp-server
uv sync --dev
source .venv/bin/activate

# Run tests
uv run pytest

# Run with MCP Inspector
mcp dev src/fetchv2_mcp_server/server.py

# Linting and type checking
uv run ruff check .
uv run pyright

License

MIT - see LICENSE for details.

Contributing

Contributions welcome! Please see CONTRIBUTING.md for guidelines.

Support

For issues and questions, use the GitHub issue tracker.