Context

The Context MCP Server provides intelligent context management and real-time web content fetching for AI assistants. It lets you store, retrieve, and organize contextual data while fetching web content, so your AI workflows stay up-to-date and well-contextualized.

How to use

You interact with Context MCP Server through an MCP client by pointing it at a local stdio server or a remote HTTP endpoint. Use the fetch tool to retrieve page content as markdown text or raw content, and use fetch_and_save to store fetched content on disk for later use. Your client can leverage these tools to gather current information, convert it to a readable format, and organize it in a clear file structure for future reference.

How to install

Prerequisites: you need Python and a way to run MCP servers (uv, uvx, or npm/yarn-based runners as shown below). Follow the concrete steps to clone, install dependencies, and start the server.

Step 1: Install the server tooling or clone the project. You have options to run via uv/uvx or to install a Python package if available.

Step 2: Start the server using a standard local runtime configuration. The following example shows how to configure Claude Desktop to run the local MCP server and specify the working data directory.

{
  "mcpServers": {
    "context_mcp_server": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/path/to/your/context-mcp-server",
        "python",
        "-m",
        "context_mcp_server"
      ],
      "env": {
        "CONTEXT_DIR": "/path/to/your/data/directory"
      }
    }
  }
}

Additional installation options

Alternative installation paths let you run the server directly from a package manager. If you prefer a package-based approach, you can install via uvx or pip and then start the server with the specified module entry point.

Below are examples that illustrate the two common local start methods you might use depending on your setup.

# Using uvx to run the local MCP server
uvx context-mcp-server

# Or install as a Python package and run the module directly
pip install context-mcp-server
python -m context_mcp_server

Configuration

Environment variables you may set to customize behavior.

CONTEXT_DIR sets the working directory where files will be saved when using the fetch_and_save tool. If not set, it defaults to data.

export CONTEXT_DIR=/path/to/your/data

Notes on usage

Remember to provide a valid directory for CONTEXT_DIR when you plan to save fetched content. You can mix use of fetch and fetch_and_save depending on whether you want to scan live pages or keep a local archive of fetched material.

Available tools

fetch

Fetches content from a URL and returns it as text, preferring the Jina Reader API with a fallback to direct HTTP request.

fetch_and_save

Fetches content from a URL and saves it to a file, using the Jina Reader API when available and falling back to direct HTTP fetch; saves to a designated file path or generates one automatically.