AllTrails MCP Server

The AllTrails MCP server is a community project by GitHub user srinath1510 that was designed to expose AllTrails hiking trail data to MCP-compatible clients like Claude Desktop. It allowed users to search for trails within a US national park and pull back detailed information about a given trail, including difficulty, length, elevation gain, ratings, route type, and descriptions.

Important: This server is DEPRECATED and non-functional as of January 25, 2026. The implementation relied on web scraping the AllTrails website with BeautifulSoup, which violated AllTrails' Terms of Service. At the request of AllTrails, all scraping functionality has been disabled. The repository remains online for historical and educational purposes only.

There is no official AllTrails MCP server. AllTrails does not currently publish a public API, and any community implementation that scrapes the site will likely run into the same ToS issues. Users who need trail data programmatically should contact AllTrails directly about authorized access.

Tools

Warning: This server is deprecated and no longer functional. Scraping has been disabled at AllTrails' request. The setup steps below reflect the historical README only.

Prerequisites:

• Python 3.8 or higher
• An MCP-compatible client such as Claude Desktop
• Unix-like environment (macOS confirmed; Linux/WSL likely compatible)

Installation (historical):

git clone https://github.com/srinath1510/alltrails-mcp-server.git
cd alltrails-mcp-server
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Claude Desktop config:

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "alltrails_mcp_server": {
      "command": "/path/to/venv/bin/python3",
      "args": ["/path/to/alltrails-mcp-server/server.py"]
    }
  }
}

No environment variables or API keys were required. A Docker option was also provided via docker build -t alltrails-mcp ..

• Historical only: Search for day hikes in a specific US national park (e.g. Great Smoky Mountains, Yosemite) by natural-language query.
• Historical only: Pull structured trail metadata (difficulty, length, elevation gain, route type) into an agent for trip planning.
• Historical only: Generate hiking recommendations filtered by attributes like waterfalls, views, or difficulty level.
• Reference: Study the repo as an example of an MCP server built around stdio transport and Python, with the caveat that the data source is now blocked.

• "What are some good day hikes in the Great Smoky Mountains?"
• "Find me a challenging trail with waterfalls in Tennessee."
• "Get the details for the Alum Cave Trail, including difficulty and elevation gain."
• "List moderate trails under 5 miles in Yosemite National Park."

• Simple, narrow scope: two clearly defined tools for trail search and trail details.
• Multiple install paths documented (venv, system Python, Docker).
• Open source and readable as a small example of a Python MCP server using stdio transport.

• Deprecated and non-functional as of January 25, 2026. Scraping was disabled at AllTrails' request after the project was found to violate their Terms of Service.
• Not an official AllTrails product. There is no official AllTrails MCP server or public API.
• Limited coverage even when working: trails had to be queried by US national park slug, not arbitrary regions or coordinates.

• OpenStreetMap / Overpass MCP servers: community-built MCP servers that query OSM data, which includes hiking trails (highway=path, route=hiking) without ToS issues.
• Google Maps MCP server (official, by Google): supports place search and routing, useful for outdoor trip planning even though it lacks AllTrails-specific trail metadata.
• National Park Service (NPS) API wrappers: a custom MCP server over the free NPS API can provide trail and park data with proper authorization.