Simple MCP Tool Server is an MCP server that provides tools for fetching web content and querying local documentation over Server-Sent Events (SSE) transport. Built using Python and the Model Context Protocol SDK, the server connects LLM clients directly to remote web pages and internal semantic documentation search indexes. Developers, researchers, and technical users utilize this server to allow AI coding assistants to extract live web content via plain HTTP requests and retrieve structured references from local Markdown documentation. It enables agents to pull raw text from any accessible URL, perform top-k vector similarity searches over the Python SDK guides, and retrieve full document contents based on search results. Operating as an ASGI application with a dedicated command-line interface, Simple MCP Tool Server manages runtime lifecycle processes such as starting, stopping, checking status, and restarting background daemon services on designated ports.
Category: Browser & Web Automation
Tags: html, http, sse, web scraping
bash python3 -m venv venv source venv/bin/activate 3. Install dependencies: bash pip install -r requirements.txt 4. Start the MCP server: bash python -m mcp_simple_tool start 5. In your project, create or update .cursor/mcp.json with the following configuration: json { "mcpServers": { "website-fetcher-sse": { "url": "http://localhost:7000/sse" } } }Part of MCP Servers
Set up a Python 3.10 or higher virtual environment, activate it, and install the project requirements using pip install -r requirements.txt. You can then manage the server process using the provided python -m mcp_simple_tool command-line interface.
The server exposes three primary tools: fetch, which downloads page text from an absolute URL; search_docs, which performs semantic searches across local SDK documentation to return top excerpts; and get_content, which retrieves full file contents for matched documentation paths.
The server operates over an SSE transport endpoint at http://localhost:7000/sse. It is documented for use with Cursor by configuring an entry in .cursor/mcp.json, but it can connect to any client supporting SSE MCP transports.
The package includes a CLI module run via python -m mcp_simple_tool. It supports start, stop, check, and restart commands, allowing you to run the server as a background daemon and configure custom ports using the --port flag.