Alaya Lab (ALab)
One-click GPU workspace with Notebook and model environments ready to go
Alaya Lab (ALab) fuses the lightweight, interactive feel of Cloud Container Instance (CCI) with the heavy compute of HyperTrain — one environment that takes you from code editing and debugging straight through to large-scale model training, with no tool-switching in between.
When to use it
- Quick model inference validation
- Dataset preprocessing and small-scale training
- Shared GPU environment for a research group
- LLM application demos and prototypes
- Promote workbench-tested code into HyperTrain for distributed training in one click
Get started
One-click activation
Spin up your first ALab from the recommended baseline
Open the workbench
Switch VSCode / Jupyter, copy SSH credentials, expose ports
Modify configuration
Compute, dev environment, env vars, storage
Create a HyperTrain job
Promote workbench code to distributed training
Choosing the right product
| Scenario | Product |
|---|---|
| Single-user interactive notebook | ALab |
| Custom containers, batch inference | CCI |
| Multi-node, Kubernetes orchestration | VKS |
| Distributed training jobs | HyperTrain |
Billing
ALab is billed per running second of the workbench; storage is billed by capacity. See the billing reference.
Last updated on
Integrate OpenClaw with Lark / Feishu
From bot creation to event callbacks — full Lark / Feishu integration for Alaya NeW OpenClaw, covering both the AlayaNeW shortcut and the SSH installation paths
One-click activation
Activate an ALab instance from the platform's recommended baseline — prerequisites, steps, exception handling