home / mcp / omnitaskagent mcp server

OmniTaskAgent MCP Server

Provides an Python-based MCP server to connect to multiple task management models and editor integrations for streamlined task management.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "acnlabs-omnitaskagent": {
      "command": "python",
      "args": [
        "run_mcp.py"
      ],
      "env": {
        "OPENAI_API_KEY": "your-key-here"
      }
    }
  }
}

You can run and integrate the OmniTaskAgent MCP server to connect your editor and development tools to a Python-based multi-model task management workflow. This MCP server enables you to start an MCP service locally, configure it in your editor, and use standardized MCP commands to manage tasks, decompositions, and analyses within your development environment.

How to use

To use the MCP service, start the local MCP server from your project root and connect your editor or LangGraph Studio to it. You will interact with the server through your editor’s MCP integration and the CLI, issuing commands to create tasks, decompose work, and manage tasks across supported models and systems. Ensure your editor is configured to communicate with the local MCP endpoint and that the required environment variables are set for your chosen AI APIs.

How to install

Prerequisites you need before installing include Python 3.11 or newer and Node.js for any optional IPC or editor tooling. Follow these steps to install and run the MCP server locally.

# Install Python package in editable mode (recommended for development)
uv pip install -e .

# Alternatively, install via pip
pip install -e .

# Install Node.js dependencies if you plan to use MCP editor integrations
npm install

Configuration

Configure the MCP service by setting the required API keys in your environment. You will typically set an OpenAI API key or an Anthropic API key depending on which models you intend to use. The MCP service also accepts optional model configuration like the default language model, creativity temperature, and maximum tokens.

# Required: API keys (configure at least one)
OPENAI_API_KEY=your_openai_api_key_here
# Or
ANTHROPIC_API_KEY=your_anthropic_api_key_here

# Optional: Model configuration
LLM_MODEL=gpt-4o  # Default model
TEMPERATURE=0.2   # Creativity parameter
MAX_TOKENS=4000   # Maximum tokens

Editor Integration (MCP Service)

To enable editor integration, run the MCP service locally and configure your editor to connect using the provided MCP settings. You can provide environment variables needed for authentication and model access through the editor’s MCP configuration.

# Start STDIO-based MCP service
python run_mcp.py

Example editor integration configuration (shown for illustration) shows how the MCP server can be declared in your editor’s settings, including the command and arguments needed to start the service.

{
  "mcpServers": {
    "task_master_agent": {
      "type": "stdio",
      "command": "+REQUIRED: command to run the MCP service+",
      "args": ["+REQUIRED: script arguments+"],
      "env": {
        "OPENAI_API_KEY": "your-key-here"
      }
    }
  }
}