After exporting a validated profile for production, you must deploy the bundle on the target host. This process is designed to work in air-gapped environments and provides flexible options for starting MCP services using Docker Compose or helper scripts.

Run the following steps to deploy the exported production bundle on the target host:

  1. Transfer the exported directory to the target host. Ensure that Docker and Docker Compose are installed and configured on the host system.

  2. Navigate to the exported bundle directory:
    cd exported/<name>
  3. Choose either of the following methods to start the MCP server:
    • Option A: Docker Compose (Preferred)

      Run the following command to start the containers in detached mode:

      docker compose -f docker-compose.mcp.yml up -d
    • Option B: Helper Script with optional overrides

      The helper script provides a simplified way to start and manage containers. For example,

      On Linux or macOS:
      ./run.sh start
./run.sh start --port 9000 --name my-server
./run.sh status   # Check container status

      On Windows using PowerShell:

      .\run.ps1 start
.\run.ps1 start -port 9000 -name my-server
.\run.ps1 status   # Check container status
    where,
    • The first command starts the MCP services using the default configuration from the exported bundle. It launches the Docker containers without any custom overrides.
    • The second command starts the MCP services with custom settings. The default port is 8500. The -port 9000 option overrides the default and sets the port to 9000, while the -name my-server option assigns a custom name to the server instance. Use this command for running multiple instances or avoiding port conflicts.
    • The third command checks the current status of the MCP containers.
    Note: Option A is the preferred method for production deployments because it provides better control and visibility.
  4. Monitor logs to verify successful startup of the containers. 
    .\run.ps1 start docker compose -f docker-compose.mcp.yml logs -f --tail=50 || \
docker-compose -f docker-compose.mcp.yml logs -f --tail=50
    Both commands are used to display and follow the logs of containers defined in the docker-compose.mcp.yml file. The first command uses the modern Docker Compose v2 syntax (docker compose), and the second uses the legacy Docker Compose CLI (docker-compose). They are combined with || so that if the first command fails, the second one runs as a fallback. Both commands show the last 50 lines of logs and continue streaming new log entries in real time.