Deploy Commvault MCP server

Connect AI systems to your Commvault environment using the Commvault MCP Server. The server provides secure access to read and manage backup jobs, storage pools, users, plans, and other supported Commvault resources. It follows the Model Context Protocol (MCP) to enable automated visibility, faster decisions, and smoother protection workflows.

The Commvault MCP Server supports the following personas:

• Backup admins: Monitor job health, check SLAs, query job history, and investigate failed jobs.

• AI integrators and developers: Connect AI tools securely to Commvault using MCP endpoints and configure integrations such as DocuSign.

• IT automation teams: Integrate Commvault workflows into other platforms such as ServiceNow.

Use cases

The following examples show common ways to use the Commvault MCP Server; actual use cases might vary by environment and integration.

• Back up a DocuSign workload: The AI assistant can configure a plan, trigger backups and restores, and set a weekly schedule.

• Check SLA health: The AI assistant can retrieve SLA data and highlight missed targets.

• Investigate failed jobs: The AI client runs job analysis and identifies failure causes through Commvault MCP.

• Create incident tickets: Use ServiceNow and Commvault MCP servers together to create incident tickets through conversational workflows.

Choose a deployment method

You can deploy the Commvault MCP Server using one of the following methods:

• Use Method 1 (Commvault hosted) when:

  • You need only Commvault-native operations (job management, storage, users, plans, schedules)

  • You want streamlined setup through Command Center

  • You don't require third-party platform integrations

• Use Method 2 (Git repository) when you want to extend the server with custom integrations such as DocuSign backup or ServiceNow ticketing integration.

Both methods can run simultaneously in the same environment if you need both Commvault-native and third-party capabilities.

Method 1: Deploy the hosted MCP server

Deploy the MCP service on a Web Server client in your Command Center environment.

1. From the navigation pane, go to Manage > Infrastructure > Servers.

2. Select the Web Server client where you want to host the service.

   If your environment has multiple Web Server clients, identify the Command Center endpoint that points to the Web Server client you select. You'll use this endpoint when connecting your AI client.

3. From the Maintenance menu, select Add software.

4. Select the MCP Service package, and then click Install.

5. Wait for the installation job to complete successfully.

   Note

   You can start using the MCP server as soon as the installation job completes successfully. The MCP Service package might not appear immediately in the Roles list.

After the package installation completes, configure your AI client to connect to the hosted service.

Configure AI clients

The MCP server can connect with various AI clients. Each client requires its own configuration file or settings entry that define how it communicates with the MCP server. For detailed integration steps, see your AI client's documentation.

The following steps explain how to connect using Visual Studio Code:

1. In Visual Studio Code, open the Command Palette (Ctrl + Shift + P).

2. Select MCP: Add MCP Server, and then choose HTTP (HTTP or Server-Sent Events).

3. Enter the MCP service URL in the following format:

   https://$[Command Center FQDN]/commandcenter/mcp

   Use the Command Center endpoint that points to the Web Server client where you installed the MCP Service package.

4. Name the server and click Start.

5. When prompted for authentication, the browser opens automatically. Sign in to the Command Center and return to Visual Studio Code.

6. Verify the connection is active.

   In Visual Studio Code, check that the MCP server displays as connected and lists the available tools.

   Note

   • The hosted service uses your Command Center credentials. No additional authentication configuration is required.

   • The Commvault-hosted MCP server provides access to all Commvault-native tools for job management, console operations, client management, storage, users, plans, and schedules.

Method 2: Deploy the MCP server from the Git repository

Set up the MCP server locally using Python and the Commvault MCP Server repository.

1. Verify that Python 3.11 or later and the uv package manager are installed on your machine.

2. To clone the repository, run:

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

3. Run the setup script:

   uv run setup.py
   

   The setup wizard guides you through selecting a transport mode (such as STDIO, Streamable HTTP, or SSE), providing connection details, and selecting an authentication method.

4. Start the server:

   uv run -m src.server
   

5. Verify the server started successfully.

Supported authentication methods

The Commvault MCP Server supports the following authentication methods.

Configure AI clients

The MCP server can connect with various AI clients. Each client requires its own configuration file or settings entry that define how it communicates with the MCP server. For detailed integration steps, see your AI client's documentation.

The following examples show Claude configurations for both remote and local setups.

Remote MCP server

{
  "mcpServers": {
    "Commvault": {
      "command": "cmd",
      "args": [
            "/c", 
            "npx", 
            "mcp-remote", 
            "$[HOST:PORT]/mcp", 
            "--header", 
            "Authorization: $[secret stored in server keyring]"
        ]
    }
  }
}

{
  "mcpServers": {
    "Commvault": {
      "command": "npx",
      "args": [
        "mcp-remote", 
        "$[HOST:PORT]/mcp", 
        "--header", 
        "Authorization: $[secret stored in server keyring]"
    ]
    }
  }
}

Local MCP server (STDIO)

{
  "mcpServers": {
    "Commvault": {
      "command": "$[/path/to/.venv]/bin/python",
      "args": ["$[/path/to]/src/server.py"]
    }
  }
}

{
  "mcpServers": {
    "Commvault": {
      "command": "$[C:\\YOUR\\PATH\\TO\\commvault-mcp-server\\.venv]\\Scripts\\python.exe",
      "args": ["$[C:\\YOUR\\PATH\\TO]\\commvault-mcp-server\\src\\server.py"]
    }
  }
}

Integrations

Extend the MCP Server with optional integrations to connect Commvault with third-party platforms.

DocuSign

Back up completed DocuSign envelopes to a Commvault S3 vault. Automate listing and restore operations directly through the AI assistant.

1. Set the following environment variable:

   ENABLE_DOCUSIGN_TOOLS=true
   

2. Configure an S3 vault in Commvault. For information about configuring an S3 vault, see Getting Started for MSP Administrators and Getting Started for Tenants.

3. In the config/ directory, perform the following:

   1. Create a docusign_config.json file using the following template.

      Template for docusign_config file

      {
          "docusign": {
              "integrationKey": "$[YOUR_INTEGRATION_KEY_HERE]",
              "userId": "$[YOUR_USER_ID_HERE]", 
              "authServer": "account-d.docusign.com",
              "scopes": "signature impersonation",
              "basePath": "https://demo.docusign.net/restapi"
          },
          "fromDate": "$[YYYY-MM-DDTHH:MM:SSZ]"
      }
      

   2. Add your DocuSign private key file docusign_key.pem.

ServiceNow

The MCP server supporting ServiceNow actions must be added to your AI client to enable the integration.

Example workflow: Create incident tickets when backup jobs fail.

1. AI client analyzes failed jobs in Commvault through MCP.

2. Identifies all failed jobs.

3. Creates ServiceNow tickets for each failure.

4. Confirms creation within your ServiceNow dashboard.