Task Manager

You run a task-focused MCP server that lets AI agents create, organize, and track tasks and projects, parse PRDs into actionable work, and generate templates. It supports both HTTP-based MCP clients and local, stdio-based workflows, making it easy to integrate into your development and planning processes.

How to use

You connect to the Task Manager MCP Server with an MCP client. Use the HTTP method when you want a remote integration and the stdio method for a local, in-process workflow. Through either path, you can create projects, add tasks, parse PRDs, expand tasks into subtasks, update task statuses, estimate complexity, and generate task templates.

Practical workflows you can perform:

• Create a new project and establish its initial task file.
• Add tasks with descriptions and subtasks to organize work.
• Convert a Product Requirements Document (PRD) into structured tasks automatically, then break them into smaller steps.
• Update the status of tasks and subtasks as you progress.
• Retrieve the next uncompleted task to focus your efforts.
• Generate file templates for new tasks and get AI-driven next-action suggestions.

How to install

Prerequisites you need before installing:

• Python 3.12+
• API keys for your chosen LLM provider (OpenAI, OpenRouter, or Ollama)
• Docker if you plan to run the server as a container (recommended)

Choose your deployment path and follow the steps below.

# Using UV (Python environment management and MCP server runtime)
pip install uv

# Install the MCP server package in editable mode from the source
uv pip install -e .

# Copy example environment and configure
cp .env.example .env

Using the Task Manager

HTTP (MCP URL) configuration for SSE transport and a local stdio configuration are both supported. Choose the path that fits your integration style.

HTTP configuration (SSE) for remote MCP client access: use the provided MCP URL endpoint to connect via SSE.

{
  "mcpServers": {
    "task-manager": {
      "transport": "sse",
      "url": "http://localhost:8050/sse"
    }
  }
}

Available tools

create_task_file

Create a new project task file to establish a workspace for task management.

add_task

Add a new task to a project with a description and optional subtasks.

update_task_status

Update the status of a task or its subtasks as work progresses.

get_next_task

Retrieve the next uncompleted task from a project to focus work.

parse_prd

Parse a Product Requirements Document into structured, actionable tasks.

expand_task

Break down a task into smaller subtasks for more granular progress tracking.

estimate_task_complexity

Estimate the complexity and time requirements for a task.

get_task_dependencies

Track and resolve dependencies between tasks to maintain flow.

generate_task_file

Generate file templates based on task descriptions to speed up setup.

suggest_next_actions

Provide AI-powered suggestions for the next steps on a task.