Machines

Outpost Machines are cloud development environments backed by GPU and CPU instances. Provision a machine in seconds, SSH in, and start developing.

Key features

• GPU selection — A10G, A100, H100, and other accelerators across AWS, Azure, and DigitalOcean.
• SSH access — connect instantly. Bring your terminal, VS Code Remote, or JetBrains Gateway.
• Multi-cloud — launch on AWS, Azure, or DigitalOcean. Switch providers without changing your workflow.
• Auto-stop — machines stop after a configurable idle timeout. You only pay for active compute time.
• Monitoring — track GPU utilization, memory, disk, and network from the dashboard.

Quick start

# Launch a machine with an A100 GPU outpost dev launch --name dev-box --cloud aws --region us-east-1 --gpus A100 # Check machine status outpost dev status # Stop when done outpost dev stop dev-box # Start it again later outpost dev start dev-box # Delete when no longer needed outpost dev delete dev-box

# Launch a machine with an A100 GPU outpost dev launch --name dev-box --cloud aws --region us-east-1 --gpus A100 # Check machine status outpost dev status # Stop when done outpost dev stop dev-box # Start it again later outpost dev start dev-box # Delete when no longer needed outpost dev delete dev-box

How it works

1. Provision — select a GPU type, cloud provider, and resource requirements. Outpost spins up the machine in under a minute.
2. Connect — SSH into the machine directly, or open it in VS Code Remote or JetBrains Gateway.
3. Develop — write code, train models, run experiments on server-grade hardware.
4. Save costs — machines auto-stop when idle. Restart in seconds when ready.

Use cases

• Model training and experimentation — iterate on training runs with GPUs without maintaining local hardware.
• Data exploration — run Jupyter notebooks on machines with direct access to cloud storage and datasets.
• Collaborative development — share a machine configuration so every engineer gets an identical environment.
• Secure development — keep source code on private cloud infrastructure rather than local laptops.

Next steps

• Create a Machine — step-by-step provisioning guide
• Connect and Develop — SSH setup, IDE integration, file sync