razorBridge

Products

GPU Blades Inference Gates Pricing

For

Universities Competitions Docs
EN |
|
|
Sign In Get Started

Quickstart

Get from zero to GPU in under 5 minutes. This guide covers installing the CLI, authenticating, joining an event, and connecting to your GPU blade.

1. Install the CLI

The rb CLI is distributed via pipx. Choose your platform:

brew install pipx
pipx install rb-cli
pip install pipx
pipx install rb-cli

Use WSL (Windows Subsystem for Linux). Then follow the Linux instructions above.

# Inside WSL terminal
pip install pipx
pipx install rb-cli

Verify the installation and enable tab completions:

$ rb --version
rb-cli 0.1.0

$ rb --install-completion

2. Sign up

Visit razorbridge.eu and create an account. You start with €10 free credit — no credit card required.

3. Authenticate

Log in from your terminal using the device auth flow:

$ rb auth login
Opening browser for authentication...
Waiting for approval...

✓ Logged in as [email protected]
  Token expires in 1 hour (auto-refreshes)

Your browser will open for you to approve the device. Once approved, the CLI is authenticated.

4. Join an event

Enter the join code your professor shared:

$ rb event join CUDA-2026
✓ Joined "ML Workshop — Spring 2026"
  GPU: L40S (48 GB VRAM)
  Starts: 2026-03-15 09:00 CET
  Duration: 3 hours
  Participants: 12/40

Join codes are case-insensitive. Your professor may share them as CUDA-2026, cuda-2026, or any variation.

5. Connect to your GPU

Once the event starts and your blade is provisioned, connect with one command:

$ rb blade ssh
Connecting to [email protected]:22012...

Welcome to your razorBridge GPU Blade
  GPU ready · CUDA + PyTorch pre-installed

student@gpu-blade:~$ nvidia-smi

rb blade ssh drops you straight onto the GPU. With sshpass installed it's fully passwordless; otherwise it prints the one-time password to paste.

To run a single command without opening a shell — handy for kicking off training or a quick check — use rb run:

$ rb run nvidia-smi
$ rb run "python train.py --epochs 10"

Prefer your editor? rb blade config --append writes an SSH alias so you can run ssh gpu-blade-abc123 — works directly with VS Code Remote-SSH, JetBrains Gateway, or any SSH client.

$ rb blade config --append
✓ Added gpu-blade-abc123 to ~/.ssh/config
  Connect with: ssh gpu-blade-abc123

6. Transfer files

Once the SSH config alias from step 5 is in place, use standard scp or rsync to move files. Upload to your blade:

$ scp ./notebook.ipynb gpu-blade-abc123:~/
notebook.ipynb                          100%   24KB   2.1MB/s   00:00

Download results:

$ scp -r gpu-blade-abc123:~/results/ ./
results/metrics.json                    100%  4.2KB   1.0MB/s   00:00
results/model.pt                        100%  152MB   88MB/s    00:01

What's next