How HCP Terraform workspaces organize a collection of infrastructure and run Terraform remotely in a shared execution environment.
An HCP Terraform workspace is the managed context for one collection of infrastructure: its configuration, variables, state, run history, and access controls. It fills a role similar to a persistent local working directory, but is a distinct feature from Terraform CLI workspaces. HashiCorp’s workspace documentation describes the differences.
With remote operations (the default execution mode), HCP Terraform runs Terraform in disposable managed workers. It queues workspace runs, records their plans and logs, and normally plans before an approved apply. Remote execution improves consistency and visibility; it does not remove the need to review a plan or scope provider permissions.
A workspace can instead use local or agent execution. In local mode, HCP Terraform can still host state while Terraform runs on a workstation; in agent mode, execution can occur in a private network. Choose execution mode for the access boundary and operational model, rather than assuming every workspace runs remotely.