local-process-intelligence

The Local Process Intelligence MCP Server provides a local monitoring assistant that leverages MCP to expose detailed insights about the host system. It can retrieve a full system summary (OS, CPU, RAM, Disk, Boot Time), monitor CPU usage in real time, track RAM and swap usage, check disk usage, analyze top processes by CPU load, search for processes by name, and explore the process tree for a given PID. It also detects anomalies such as memory leaks (high memory usage with low CPU), zombie/defunct processes, and resource usage spikes, delivering actionable warnings and insights suitable for driving a local agent or Claude Desktop integration. To use it with MCP-enabled tooling, run the server and connect your MCP client or agent to the configured endpoint to start receiving structured observations, summaries, and diagnostics.

Usage scenarios include running a local agent that can query the server for diagnostic data, feed it into a reasoning loop, and obtain concrete suggestions for optimizing system performance or addressing anomalies in real time.

Prerequisites:

1. Have Python installed on your system (3.8+ recommended).
2. Install UV (for MCP Py SDK guidance, though this server can be run directly with Python).
3. Ensure you have access to the repository containing this MCP server (clone or download).

Install and run:

# 1. Clone the repository (if you haven’t already)
git clone <repository-url>
cd <repository-directory>

# 2. (Optional) Create and activate a virtual environment
python -m venv venv
source venv/bin/activate  # on macOS/Linux
# Windows: .\venv\Scripts\activate

# 3. Install dependencies (if any are specified in requirements.txt or setup.py)
pip install -r requirements.txt  # if provided

# 4. Run the MCP server entry point (as configured in mcp_config)
# Depending on your setup, this could be executing main.py directly or via UV/MCP tooling
python main.py

Note: If you use the MCP tooling with UV, you can also initialize and install the server via the MCP commands documented in the quickstart (e.g., uv sync, uv run mcp install main.py, or mcp install main.py). Adjust commands if your environment uses a virtual environment or containerized setup.

Tips and notes:

• Ensure Python is accessible from your shell (PATH configured).
• If you modify environment variables, keep PYTHONUNBUFFERED=1 to ensure live logs for MCP observers.
• When integrating with Claude Desktop or other MCP clients, restart the app after installing or updating the server, as noted in the Quickstart.
• If you encounter permissions issues on macOS, run with elevated privileges or adjust your security settings to allow script execution.
• The server is designed for local, macOS-focused environments per the compatibility notes; cross-platform behavior may vary on Linux/Windows.