Hosted cloud Linux desktop that any MCP client drives: an agent gets Chrome, a shell and a full Ubuntu desktop to browse, click, type, read the screen and run commands.
Use it now
Keep it with up to 9 other tools you're setting up, then share them all as one link.
Try this first
Ask your assistant, with box by Simular connected:
- Run Shell Commands
Use box by Simular to run commands and scripts in the box's persistent shell: ‹what you need›
- Keep API Keys Hidden
Use box by Simular to store API keys as box secrets: ‹what you need›
- Watch and Take Control
Use box by Simular to watch the agent's desktop live in the browser: ‹what you need›
Connect it over MCP
{
"mcpServers": {
"simular-box": {
"url": "https://mcp.box.simular.ai/mcp"
}
}
}claude mcp add --transport http simular-box https://mcp.box.simular.ai/mcp
In Grok, ChatGPT and other assistants, add a custom connector with the address https://mcp.box.simular.ai/mcp (how, step by step).
It may ask you to sign in to Simular or for a key when you add it; the setup guide says which.
Setup guide: box.simular.ai/install
Install the CLI
npx @simular-ai/box@1.2.1 connect
Setup guide: box.simular.ai/install
Compare
Press Compare to add it to the tray at the bottom of the window, then pick up to 2 others anywhere on the site.
- Access
- MCPCLI
- Auth
- OAuth 2.0
- Skills
- 4
Skills
Use a Cloud Desktop
Give an agent a persistent Ubuntu desktop with Chrome to open sites, click, type and read the screen from any MCP client.
Run Shell Commands
Run commands and scripts in the box's persistent shell, with files and state kept while the desktop sleeps.
Keep API Keys Hidden
Store API keys as box secrets so commands use the real key while the AI only ever sees a stand-in value.
Watch and Take Control
Watch the agent's desktop live in the browser and take over the mouse and keyboard when a human is needed.