🔑 SSH Access Guide

Connect to your GPU instance in 3 simple steps.
Choose your operating system below.

1

Download cloudflared for Windows

Download the cloudflared client — a small tool that creates a secure tunnel to your GPU instance.

⬇ Download cloudflared.exe
💡 Save the file as cloudflared.exe in your Downloads folder or C:\Windows\System32 for easy access.
2

Open PowerShell or Command Prompt

Press Win + X and select Windows PowerShell or Terminal.

3

Connect to your GPU instance

Replace gpu52 with your instance username shown on the instances page.

ssh -o PubkeyAuthentication=no -o ProxyCommand="cloudflared.exe access ssh --hostname bastion.opensource-ai-cloud.uk" [email protected]
💡 Password: gpu123 — type it when prompted and press Enter.

📁 File Transfer Guide

Upload scripts, datasets and download results from your GPU instance

1

Upload a file to your GPU instance

Open PowerShell and use scp with cloudflared to upload files from your PC to the GPU.

scp -o ProxyCommand="cloudflared.exe access ssh --hostname bastion.opensource-ai-cloud.uk" C:\Users\YourName\script.py gpu{ID}@bastion.opensource-ai-cloud.uk:/root/
💡 Replace C:\Users\YourName\script.py with your actual file path and {ID} with your instance number. Password: gpu123
2

Upload an entire folder

Use scp -r to upload a whole project folder recursively.

scp -r -o ProxyCommand="cloudflared.exe access ssh --hostname bastion.opensource-ai-cloud.uk" C:\Users\YourName\my-project gpu{ID}@bastion.opensource-ai-cloud.uk:/root/
3

Download results back to your PC

After training or processing, download outputs from your GPU instance back to your Windows PC.

scp -o ProxyCommand="cloudflared.exe access ssh --hostname bastion.opensource-ai-cloud.uk" gpu{ID}@bastion.opensource-ai-cloud.uk:/root/output.pt C:\Users\YourName\Downloads\
4

Use VS Code Remote SSH (recommended)

The easiest way to work on your GPU instance from Windows — edit files, run code, and drag-drop files all from VS Code.

1. Install VS Code and the Remote SSH extension

2. Add this to your SSH config file C:\Users\YourName\.ssh\config:

Host gpu-instance   HostName bastion.opensource-ai-cloud.uk   User gpu{ID}   ProxyCommand cloudflared.exe access ssh --hostname bastion.opensource-ai-cloud.uk

3. In VS Code press Ctrl+Shift+P → Remote-SSH: Connect to Host → select gpu-instance

💡 Pro Tips

Install packages inside SSH

pip install torch transformers

Run script in background

nohup python3 train.py & disown

Check GPU usage

nvidia-smi

Clone from GitHub

git clone https://github.com/you/repo

Keep session alive

tmux new -s training

Check disk space

df -h