mcpo-control-panel 0.0.4

Web UI and API for managing MCP Orchestrator (mcpo) instances and configurations.

MCPO Control Panel

A web-based control panel designed to simplify the management of MCP-to-OpenAPI (mcpo) instances and their server configurations. It provides a user-friendly UI and an API for interacting with mcpo.

Demo

https://github.com/user-attachments/assets/dc3f11de-82f6-42ee-a72f-7181c9af0f45

Key Features

• Server Definition Management:
  • Create, Read, Update, and Delete server definitions (for stdio, sse, streamable_http types).
  • Easily toggle servers as enabled/disabled for inclusion in the mcpo configuration.
  • Intuitive forms for specifying commands, arguments, environment variables, and URLs.
• Bulk Server Operations:
  • Add multiple server definitions at once by pasting JSON content.
  • Supports various JSON formats (object with mcpServers, direct name-to-config map, list of server objects).
  • Two-step bulk add process:
    1. Analyze: Preview valid new servers, existing ones, and invalid entries from your JSON.
    2. Confirm: Add only the validated new servers to the database.
  • Automatic de-adaptation of Windows-specific commands (e.g., cmd /c npx ... -> npx ...).
• MCPO Process Control:
  • Start, stop, and restart the mcpo process directly from the UI.
  • View the real-time status of the mcpo process (Running, Stopped, Error).
  • "Apply and Restart" functionality: Generates the mcpo configuration from current server definitions and then restarts mcpo.
• Log Viewing:
  • Display the latest logs from the mcpo process log file.
  • Configurable auto-refresh for logs.
  • Manual refresh and auto-scroll options.
• MCPO Settings Configuration:
  • Manage mcpo startup parameters (port, API key usage, config file path, log file path).
  • Configure UI behavior (log refresh interval).
  • Setup and manage health check parameters for mcpo.
• Tool Aggregation & Viewing:
  • Dynamically fetches and displays a list of available tools (server paths and summaries) from a running mcpo instance by querying its OpenAPI specifications.
  • Provides the base URL for constructing tool invocation links.
• Health Checks & Auto-Restart:
  • Background health monitoring of the mcpo process using a built-in echo server.
  • Configurable check interval, failure attempts, and retry delays.
  • Optional automatic restart of mcpo if health checks fail consecutively.
• Dynamic Configuration Generation:
  • Generates the mcp_generated_config.json file for mcpo based on enabled server definitions.
  • Provides options to download the standard configuration or a Windows-adapted version (which wraps commands like npx, uvx, docker with cmd /c ...).
• Modern Web UI:
  • Built with FastAPI, Jinja2 templates, and HTMX for a responsive and dynamic user experience.
  • Utilizes Materialize CSS for styling.
  • Persistent storage using an SQLite database (via SQLModel).

Installation

Prerequisites

• Python 3.11 or higher.

Method 1: From PyPI (Recommended)

pip install mcpo-control-panel

Method 2: From Source (for development)

1. Clone the repository:

   git clone https://github.com/daswer123/mcpo-control-panel.git
   cd mcpo-control-panel
   

2. Install in editable mode (this will also install dependencies):

   pip install -e .
   

Running the Application

Once installed, you can run the MCPO Control Panel using the following command:

python -m mcpo_control_panel [OPTIONS]

Or, if you have the scripts directory of your Python environment in your PATH:

mcpo-control-panel [OPTIONS]

Common Options:

• --host TEXT: Host to bind the server to. (Default: 127.0.0.1)
  • Environment variable: MCPO_MANAGER_HOST
• --port INTEGER: Port to bind the server to. (Default: 8083)
  • Environment variable: MCPO_MANAGER_PORT
• --workers INTEGER: Number of Uvicorn workers. (Default: 1)
  • Environment variable: MCPO_MANAGER_WORKERS
• --reload: Enable auto-reload (for development).
• --config-dir TEXT: Directory for storing MCPO manager data (SQLite database, PID files, generated configs, settings).
  • Default: ~/.mcpo_manager_data (e.g., C:\Users\YourUser\.mcpo_manager_data on Windows or /home/youruser/.mcpo_manager_data on Linux).
  • Environment variable: MCPO_MANAGER_DATA_DIR

Example:

python -m mcpo_control_panel --port 8083 --reload --config-dir "/path/to/my/mcpo-data"

Or using environment variables:

export MCPO_MANAGER_PORT=8083
export MCPO_MANAGER_DATA_DIR="/path/to/my/mcpo-data"
python -m mcpo_control_panel --reload

The application will create the specified --config-dir if it doesn't exist. All application data, including the SQLite database (mcp_manager_data.db), settings file (mcpo_manager_settings.json), generated MCPO configuration (mcp_generated_config.json), and the mcpo process PID file, will be stored in this directory.

Accessing the UI

Once the server is running, open your web browser and navigate to:

http://<host>:<port>/ui

For example, if running with default settings: http://127.0.0.1:8083/ui

Configuration

Most configurations for the mcpo process itself (like its port, whether it uses an API key, log file paths) and for the UI (like log auto-refresh settings, health check parameters) are managed through the web interface on the "MCPO Settings" page.

These settings are saved to mcpo_manager_settings.json within the MCPO_MANAGER_DATA_DIR_EFFECTIVE (the directory specified by --config-dir or its default).

License

This project is licensed under the MIT License