You can call APIs all day, but sometimes you need a real desktop. The /v1/machines API lets you provision cloud VMs that run Windows, macOS, or Linux. Your computer use agent can then drive that desktop, browser, or terminal just like a human. No more brittle selectors or broken selectors. You get a real screen to look at and a cursor to move.
How it works
The flow is simple. You POST to /v1/machines with a machine_id and optional specs. The service provisions a cloud VM and returns its status. Once the machine is running, your agent uses the /v1/runs endpoints to drive the desktop using a computer use agent. The machine ID is your handle throughout the session.
# Provision a cloud machine
export COASTY_API_KEY=$(cat ~/.coasty/key)
curl -X POST https://coasty.ai/v1/machines \
-H "X-API-Key: $COASTY_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"machine_id": "my-desktop-01",
"os": "windows",
"screen_width": 1920,
"screen_height": 1080
}'Provisioning fields
- machine_id: required, unique identifier for the machine.
- os: either windows, macOS, or linux.
- screen_width: integer, recommended 1920 or higher.
- screen_height: integer, recommended 1080 or higher.
- The API returns a status object with machine_id and state (pending, running, stopped).
Once the machine state is running, use /v1/runs to drive it with a computer use agent.
Where this beats brittle automation
Most automation tools rely on XPath, CSS selectors, or hardcoded coordinates. If a UI changes, those selectors break. A computer use agent, powered by the Coasty API, sees the screen and acts like a human. It can adapt to layout shifts, dynamic elements, and unfamiliar interfaces. You get higher success rates on real workflows and less maintenance.
Use /v1/machines to spawn real desktop environments on demand. Pair it with /v1/runs to drive agents that see and click like a human. Build reliable, screen-aware automation at scale. Get an API key at https://coasty.ai/developers and start provisioning machines today.
Want to see this in action?
View Case Studies