Skip to main content
OpenHive can execute a colony on your computer or on a cloud computer. Choose based on where the task’s files, browser session, and connected services are available.
Screenshots show the desktop interface with fictional example data. Names, messages, counts, and connection states are illustrative, not account defaults or evidence of live activity.
Check the Cloud section to understand the colony's execution location.

Check the Cloud section to understand the colony's execution location.

​
Deploy a colony

  1. Run a small task locally and confirm its output.
  2. Open the colony’s Automation tab.
  3. In Cloud, select Deploy.
  4. Read the confirmation dialog and confirm the deployment when you are ready.
  5. Wait for upload and deployment to finish. Check that the section reports Running on a cloud computer.
  6. Choose Open to inspect the cloud environment.
Resolve deployment errors before relying on the colony to continue unattended. The availability of cloud computers and upgrades depends on the options offered to your account.
Review the colony and confirmation before running it in the cloud.

Review the colony and confirmation before running it in the cloud.

​
Check access after deployment

A cloud computer is a separate execution environment. Check that the task can access its required files, connected accounts, and browser pages there. Do not assume a website signed in on your laptop is also signed in on the cloud computer. Start with a small verification task before enabling a recurring workflow. For browser automation, inspect the cloud browser directly when login or an interactive step is needed.

​
Move work back to your computer

The Cloud section exposes advanced controls, including moving the colony back to local execution. After moving it back, keep the desktop runtime available and verify the next run locally.

​
Understand the data boundary

Desktop execution does not make OpenHive an offline-only product. Cloud services handle account and workspace functions; connected-account credentials can sync to the local runtime; model providers and third-party tools receive the inputs needed to fulfill requests. Choose the connected account deliberately, and keep private information out of a task when it is not needed. For access setup, see Credentials.