You can run an MCP server that lets AI agents control virtual computers through Orgo. This guide shows how to connect to a remote MCP endpoint or self-host a local server, and how to start using it to manage virtual machines, run commands, take screenshots, and automate tasks.

How to use

Connect to the Orgo MCP server from your MCP client to start managing virtual computers. You can use a remote MCP endpoint to access the service, or run a local server and connect to it from your client. Once connected, you can create virtual computers, perform shell commands, take screenshots, upload and download files, stream activity, and interact with AI models through the provided tools.

How to install

Prerequisites: ensure you have a modern Python runtime installed for local self-hosting, or a capable MCP client installed (for example Claude MCP client) to connect to a remote endpoint.

Option 1 — Use the remote MCP URL with your client (no local server required):

# Configure your MCP client to point at the remote endpoint
# Example for a CLI-based client:
claude mcp add --transport http orgo https://orgo-mcp.onrender.com/mcp \
  --header "X-Orgo-API-Key: YOUR_API_KEY"

Option 2 — Self-host a local HTTP server using Python (stdio transport):

# Install the package in editable mode
pip install -e .

# Run the local MCP server (stdio transport)
MCP_TRANSPORT=http python orgo_mcp.py

This will start a local server accessible at http://localhost:8000/mcp when using the default settings. You can configure your client to connect to this local endpoint.

Option 3 — Run the server in a Docker container (production-ready):

Build the image

docker build -t orgo-mcp .

Run the container

docker run -p 8000:8000 orgo-mcp

This launches the MCP server and exposes the MCP endpoint at http://localhost:8000/mcp.

Option 4 — Deploy to Render.com (cloud):

# Follow the deployment flow in the Render dashboard to create a new blueprint using your GitHub repository, then deploy.
# The server will be available at a URL like https://orgo-mcp-xxxx.onrender.com

Note: Use the remote endpoint URL shown after deployment when configuring your MCP client.

Option 5 — Deploy to Fly.io (alternative cloud):

Install flyctl, then log in and deploy

curl -L https://fly.io/install.sh | sh fly auth login fly launch --no-deploy fly deploy

Your server will be at a URL similar to https://orgo-mcp.fly.dev.

Configuration and security notes

Always keep your API key private and never expose it in public code or logs. When using a remote MCP URL, include your key in the request header as shown in your client configuration. If you run the server locally, set the ORGO_API_KEY environment variable to a secure value.

Environment variables you may configure include MCP_TRANSPORT, MCP_HOST, MCP_PORT, ORGO_API_KEY, and CORS_ORIGINS. Use these to tailor how the server listens and which origins can access it.

Examples of common usage patterns

Create and manage virtual computers, execute shell commands, take screenshots, and transfer files between your client and the MCP server. You can also start and stop streams, and run AI model interactions through the provided tools.

Troubleshooting and quick checks

If you encounter a connection issue, verify that the MCP endpoint is reachable and that the API key is correctly passed in requests. For local self-hosts, ensure the server is running and listening on the configured port. Check firewall and proxy settings if problems persist.

Notes on self-hosting and deployment

Self-hosting options include a local Python implementation, a Docker container, and cloud deployments. Depending on your setup, you may also configure a local or remote HTTP transport for client connections and supply the necessary environment variables.

Available tools

orgo_list_projects

List all projects in the Orgo MCP environment.

orgo_create_project

Create a new project in the Orgo MCP workspace.

orgo_get_project

Retrieve details for a specific project.

orgo_delete_project

Delete a project from the workspace.

orgo_start_project

Start a project, initializing related resources.

orgo_stop_project

Stop a running project.

orgo_restart_project

Restart a project to apply changes.

orgo_list_computers

List all virtual computers available to the user.

orgo_create_computer

Create a new virtual computer with specified specs.

orgo_get_computer

Get details for a particular computer.

orgo_start_computer

Power on a virtual computer.

orgo_stop_computer

Power off a virtual computer.

orgo_restart_computer

Restart a running computer.

orgo_delete_computer

Remove a computer from the environment.

orgo_screenshot

Capture a screenshot from a specified computer.

orgo_click

Simulate a mouse click at given coordinates on the computer screen.

orgo_double_click

Simulate a double-click action on the screen.

orgo_type

Type text into the active input field on the computer.

orgo_key

Send a single key press to the computer.

orgo_scroll

Scroll the viewport of the computer screen.

orgo_drag

Drag a cursor from one coordinate to another.

orgo_wait

Wait for a specified duration or until a condition is met.

orgo_bash

Open a shell on the target computer and run commands.

orgo_exec

Execute arbitrary commands on the computer.

orgo_list_files

List files in a directory on the target computer.

orgo_upload_file

Upload a file to the target computer.

orgo_export_file

Export a file from the target computer.

orgo_download_file

Download a file from the target computer.

orgo_delete_file

Delete a file on the target computer.

orgo_start_stream

Begin streaming the computer's display or session.

orgo_stream_status

Check the status of an ongoing stream.

orgo_stop_stream

Stop an active stream.

orgo_list_ai_models

List available AI models for completion tasks.

orgo_ai_completion

Run an AI completion against a chosen model.