commvault

The Commvault MCP Server provides a Model Context Protocol (MCP) interface for AI agents and automation tools to securely access and manage Commvault environments. It centralizes access to job details, commcell metrics, client and storage information, user permissions, plan configurations, and backup schedules, enabling agents to query, monitor, and control resources as needed. The server supports both OAuth-based authentication and traditional token-based authentication, with setup guiding you through the appropriate credentials and secure storage options. Once running, clients can connect to the MCP endpoint and issue structured requests to retrieve or manipulate Commvault data, automating workflows such as discovering backup schedules, inspecting job histories, and monitoring storage utilization. When using OAuth, ensure your provider is configured correctly and the redirect URI is set to the MCP base URL. For token-based setups, tokens and a secret key are securely stored or provided during setup, and the server will enforce token-based access via the MCP client header. The server can be run locally for development or deployed behind a TLS-enabled reverse proxy for production security.

Prerequisites:

• Python 3.11 or higher
• uv package manager (install via pipx, pipx is recommended; uv is used to run the server)
• Git (to clone the repository)

1. Clone the repository

git clone https://github.com/Commvault/commvault-mcp-server.git
cd commvault-mcp-server

2. Set up a Python environment Option A: Virtual environment (recommended)

python3 -m venv .venv
source .venv/bin/activate  # on macOS/Linux
# or on Windows
.\.venv\Scripts\activate

Option B: System Python (if you prefer)

3. Install uv and project dependencies

# Ensure you have uv available (via pipx or pip). Example with pip:
pip install uv
# If a requirements file exists, install dependencies there
pip install -r requirements.txt  # if provided

4. Configure the server

• Follow the Setup script to configure transport mode, server connection details, and authentication (OAuth or token-based).
• From the repository root, run the setup wizard:

uv run setup.py

5. Start the MCP server

uv run -m src.server

6. Verify local access

• By default, connect to the MCP endpoint exposed by the server (e.g., http://127.0.0.1:8000/mcp or as configured in your env).

Notes:

• If you plan to expose the MCP server in production, configure a reverse proxy with TLS (e.g., Nginx or Caddy) and restrict access to localhost or trusted networks.

Tips and common issues:

• Ensure Python 3.11+ is used and that uv is installed for efficient asynchronous operation.
• If you choose token-based authentication, you will store sensitive credentials in a secure backend. Ensure the chosen OS-native keyring is available (Windows Credential Manager, macOS Keychain, or Linux Secret Service / KWallet).
• When using OAuth, verify the OAuth provider configuration and ensure the redirect URI matches OAUTH_BASE_URL/auth/callback.
• For local development, bind to localhost (127.0.0.1) to avoid exposing the MCP server publicly.
• If you see connectivity issues from clients, confirm MCP_HOST and MCP_BASE_URL settings, and ensure the reverse proxy (if used) is forwarding the correct path.
• The MCP configuration supports both remote (streamable-http / SSE) and local STDIO modes; choose the method that matches your deployment scenario.