Welcome to the Remote Docker Extension! This guide will help you get started with managing remote Docker environments.
Install the extension from Docker Desktop or using the command line:
docker extension install telkombe/remote-docker:latest- Open the Remote Docker extension in Docker Desktop
- Navigate to the "Environments" tab
- Click "Add Environment"
- Enter the following details:
- Name: A friendly name for your environment
- Host: SSH connection string (e.g.,
ssh://user@hostname) - SSH Key: Path to your SSH private key
- Go to the "Environments" tab
- Click on the environment you want to connect to
- The extension will establish an SSH tunnel to the remote Docker daemon
Once connected, you can:
- View all containers on the remote host
- Start/Stop/Restart containers
- View container logs
- Remove containers
The Images tab allows you to:
- View all images on the remote host
- Pull new images
- Remove unused images
- View image details
In the Volumes tab, you can:
- List all volumes
- Create new volumes
- Remove unused volumes
- View volume details
The Networks tab provides:
- List of all networks
- Network details and configuration
- Create/remove networks
The MCP Servers tab allows you to run Model Context Protocol servers on remote Docker hosts, enabling AI assistants to interact with your remote systems.
- Navigate to the "MCP Servers" tab
- Click the "Browse Catalog" tab
- Browse or search for MCP servers:
- Use the search bar to find servers by name or description
- Filter by category (storage, docker, shell, kubernetes, etc.)
- View download counts and ratings
- Click "Install" on any server
- Optionally provide a custom name
- Click "Install & Start" to deploy
- Navigate to the "MCP Servers" tab
- Stay on the "Installed Servers" tab
- Click "New Server"
- Choose from pre-configured templates:
- Filesystem Access: Browse and edit files on the remote host
- Docker Management: Control Docker via MCP
- Shell Access: Execute bash commands
- Kubernetes: Manage K8s clusters
- Database Access: PostgreSQL, MySQL, MongoDB
- Git Repository: Work with Git repos
- Monitoring: System metrics and logs
- AI/LLM Integration: Connect to AI models
- Optionally customize the server name
- Click "Create" to deploy the server
- Start/Stop: Use the play/stop buttons to control server state
- Delete: Remove servers when no longer needed (must be stopped first)
- Status Monitoring: View real-time server status
- Port Information: Each server gets a unique port for connections
Once an MCP server is running:
- Note the assigned port number
- Configure your AI assistant to connect via SSH tunnel
- The MCP server will be accessible at
localhost:[port]through the tunnel
- All MCP connections are tunneled through SSH
- Each server runs in an isolated Docker container
- Filesystem access can be restricted to specific directories
- Shell commands can be whitelisted or blacklisted
- No direct port exposure to the internet
Enable auto-refresh to automatically update the displayed information:
- Click the refresh icon in the top-right
- Select your preferred refresh interval
To view container logs:
- Go to the Containers tab
- Click on a container name
- View real-time logs with filtering options
Quickly switch between environments:
- Use the environment selector in the header
- Click "Disconnect" to disconnect from current environment
If you cannot connect to a remote environment:
- Verify SSH connectivity:
ssh user@host - Check that Docker is running on the remote host
- Ensure your SSH key has proper permissions (600)
If you see permission errors:
- Ensure your user is in the docker group on the remote host
- Try connecting with
sudoif needed
Ctrl/Cmd + R- Refresh current viewCtrl/Cmd + K- Quick searchEsc- Close dialogs