Workspaces
A workspace is a cloud VM attached to a project. It provides a full Linux environment (Ubuntu 22.04) with persistent storage, terminal access, and the AI agent runtime.Workspace States
Workspaces follow a defined lifecycle:| State | Description |
|---|---|
stopped | VM is off, persistent volume is preserved |
starting | VM is being provisioned or resumed |
running | VM is active and accepting connections |
stopping | VM is shutting down gracefully |
error | VM failed to start or hit a fatal error |
stop_failed | VM failed to stop cleanly |
Workspaces automatically stop after an idle timeout (default: 10 minutes with no active connections or tasks). This prevents unnecessary resource usage.
Starting a Workspace
- Web
- CLI
Open a project and click Start Workspace. The UI shows a real-time status indicator as the workspace boots.
Stopping a Workspace
Listing Workspaces
Getting Workspace Details
Deleting a Workspace
Hardware Presets
Each workspace runs on configurable hardware. You can set these when creating a workspace or update them later.| Parameter | Description | Range |
|---|---|---|
cpu_kind | CPU type | shared or performance |
cpus | Number of CPU cores | 1 - 8 |
memory_mb | Memory in megabytes | 256 - 16384 |
volume_size_gb | Persistent storage in GB | Configurable per plan |
gpu_kind | Optional GPU type | Depends on availability |
SSH Access
You can SSH directly into a running workspace for manual debugging or exploration:SSH access requires the workspace to be in
running state. If the workspace is stopped, start it first with aether workspace start.