Skip to content
openlaunch

Your agents. Your devices.

Hardware, Linux computers, and your Home Assistant home.
Connected to the agents you already use.

Get started

Keep it on your own machine? CLI, SDK, or self-hosted

Start with something useful.

Connect a computer, bring your home into the loop, or add a device of your own.

Ask for what you actually need.

Your agent reads the device’s functions, requests an action, and follows its receipt.

01 / Linux
“Find my Pi’s specs, then open YouTube on its desktop.”

System information, browser launch, and a screenshot to see the result.

See Linux functions
02 / Home Assistant
“Turn on the test toggle, read its state, then turn it off.”

A real service call and state read. Start with a helper; no hardware needed.

Try the test helper
03 / Your hardware
“Show ‘hello’ on my Uno, then check its health.”

Built-in functions today. Add your own handlers as your project grows.

Connect an Uno

Example requests. Available functions depend on your adapter and saved grants. See what’s been verified →

Put your devices in the flow.

One device or a whole Home Assistant installation. Your agent uses the same connection.

Choose the access. Grant individual functions, choose an expiry, and revoke access from your console.

Follow the result. See what was queued, delivered and reported. Device state and physical verification stay distinct.

Use the interface that fits.

Your hardware.
Your functions.

Describe the inputs. Write the handler. Connect your agent.

Adapters publish the functions they actually implement. Your agent discovers the current schema through MCP, API or CLI, with the same permission checks across all three.

Start with one device.

Free hosted access. MIT licensed. Self-host it when you want.

Open your console See pricing →Latest updates →