FAQ

Common questions about Orgo.

What is Orgo?

Orgo provides instant desktop environments for AI agents to control computers. Think full Ubuntu or Windows machines that models like Claude, GPT, and Gemini can see and interact with through screenshots, mouse, and keyboard.

Read the docs

How is this different from headless browsers?

Headless browsers automate web pages through DOM manipulation. Orgo provides full desktop environments where AI agents see and control the screen like humans do—enabling interaction with any application, not just websites.

What latency should I expect?

Desktop boot time is ~500ms. For a full action loop (screenshot → model inference → action), expect 2-4 seconds depending on the model. Network latency to Orgo adds minimal overhead, typically under 50ms.

Which AI models are supported?

Any model with computer use capabilities. Tested with Anthropic Claude (3.5, 3.7, 4), OpenAI GPT-4o, and Google Gemini 2.5. The platform is model-agnostic.

View integration guides

What operating systems are available?

Linux (Ubuntu 24) is generally available. Windows is in private beta.

Can agents install software?

Yes. Agents have full control and can install packages via apt, pip, npm, or any package manager. Installed software persists during the session but resets on termination.

What does it cost?

Free tier includes 10 compute hours/month with no credit card required. Paid plans start at $20/month for 300 hours.

View pricing

Can I self-host?

Orgo is a hosted platform. For local development, you can use Anthropic's Docker container. For self-hosted enterprise deployments, contact us.

Contact sales

API or dashboard—which should I use?

Start with the dashboard (playground) for testing and debugging with live screenshots. Move to the API for production agents that need to scale or integrate into your application.