home / mcp / software planning mcp server

Software Planning MCP Server

Provides an MCP server for software planning goals, implementation plans, and to-dos with SSE or stdio transport.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "wang-guanye-software-planning-mcp": {
      "url": "http://localhost:8000/sse",
      "headers": {
        "SOFTWARE_PLANNING_HOST": "0.0.0.0",
        "SOFTWARE_PLANNING_PORT": "8000",
        "SOFTWARE_PLANNING_DEBUG": "false",
        "SOFTWARE_PLANNING_TRANSPORT": "sse"
      }
    }
  }
}

You have a Python-based MCP server that helps you plan software development, manage to-dos, and track task progress. It exposes a structured MCP interface you can connect to from an MCP client like Cursor, enabling you to start planning sessions, add to-dos, and save plans directly from your workflow.

How to use

Establish a connection to the MCP server from your MCP client and start planning sessions that organize your software development goals into actionable steps.

Typical usage pattern focuses on creating a planning session, outlining a goal, and progressively adding to-dos as your plan evolves. You can check the current goal and the planned tasks, update their completion status, and save the finalized plan for reference.

How to install

Prerequisites you need before installation: you must have Python installed on your system and a shell environment to run commands.

# Clone the project repository
git clone https://github.com/Jo-Lion/software-planning-mcp.git
cd software-planning-mcp

# Create and activate a virtual environment
python -m venv .venv

# Activate the virtual environment
# Linux / macOS
source .venv/bin/activate
# Windows
.venv\Scripts\activate

# Install dependencies in editable mode
pip install -e .

To run the server, you can choose between the SSE transport for Cursor integration or the stdio transport for command-line testing.

# Start with SSE transport (recommended for Cursor integration)
python server.py --debug

# Start on a custom port
python server.py --port 9000 --debug

# Start on a custom host
python server.py --host 127.0.0.1 --debug

# Start with stdio transport for command-line testing
python server.py --transport stdio --debug

Additional notes

Configure your MCP client (Cursor) to connect to the server at the SSE URL. If you run the server on the default port and host, the MCP URL is http://localhost:8000/sse. Adjust the URL if you started the server with a different port or host.

Environment variables that control server behavior include port, host, transport mode, and debug mode. You can set these on your system or via your startup script to tailor how the server runs.

Available tools

start_planning

Begin a new planning session and set the overall software development goal.

save_plan

Persist the current implementation plan for future reference.

add_todo

Add a new task to the current plan with title, description, complexity, and optional code example.

remove_todo

Remove a task from the current plan by its identifier.

get_todos

Retrieve all tasks in the current planning session.

update_todo_status

Mark a task as complete or incomplete by its identifier.