Quick Start
Get up and running in 2 minutes.
Prerequisites
You need an SSH key. If you don't have one:
ssh-keygen -t ed25519
1. Connect to a container
SSH to runtime.ovh with any workspace name as the username:
ssh myproject@runtime.ovh
You'll be dropped into an Alpine Linux container. Your workspace is mounted at /data/.
2. Upload some code
Open a new terminal and use SCP to upload files:
# Create a simple Python script
echo 'print("Hello from runtime.ovh!")' > hello.py
# Upload it
scp hello.py myproject@runtime.ovh:/data/
3. Run your code
Back in your SSH session, or connect with Python:
ssh myproject:python@runtime.ovh
cd /data
python hello.py
Tip: The :python suffix tells runtime.ovh to use the Python container image instead of Alpine.
4. Your files persist
Disconnect and reconnect later - your files are still there:
# Disconnect (Ctrl+D or exit)
exit
# Reconnect anytime
ssh myproject@runtime.ovh
ls /data/ # hello.py is still there!
Available images
| Image | Usage | Includes |
|---|---|---|
alpine |
ssh ws@runtime.ovh |
Minimal Linux, busybox |
python |
ssh ws:python@runtime.ovh |
Python 3.11, pip |
go |
ssh ws:go@runtime.ovh |
Go 1.21, git |
node |
ssh ws:node@runtime.ovh |
Node 20, npm |
rust |
ssh ws:rust@runtime.ovh |
Rust, cargo |
ubuntu |
ssh ws:ubuntu@runtime.ovh |
Ubuntu 22.04, apt |
Execute commands directly
You can run commands without an interactive shell:
# Run a command and exit
ssh myproject:python@runtime.ovh "python /data/hello.py"
# Chain commands
ssh myproject:go@runtime.ovh "cd /data && go build ./... && ./myapp"
Run background jobs
Use the run prefix to start jobs that continue after you disconnect:
# Start a long-running job
ssh myproject:python@runtime.ovh run "python train.py --epochs 100"
# Monitor running jobs
ssh cli@runtime.ovh ps
# Connect to a running job
ssh session-abc123@runtime.ovh
Get help
View all available commands and options for your workspace:
ssh myproject@runtime.ovh help
Next steps
- Learn about workspaces - Organize your projects
- GPU access - Use V100, A100, H100 for ML
- CLI commands - Manage images, aliases, API keys
- AI Agents - Use the built-in shai agent