Getting StartedQuickstart (60 Seconds)

Quickstart — GPU in 60 Seconds

Get a GPU instance running in under a minute.

Using the Dashboard

  1. Sign up at app.pulserun.dev
  2. Add $5+ credit via Stripe
  3. Click Launch Instance
  4. Select GPU type (e.g., A100 80GB)
  5. Pick a template (e.g., PyTorch)
  6. Click Launch — done!

Your instance will be ready in ~30 seconds with SSH access and Jupyter URL.

Using the API

# Set your API key
export PULSERUN_API_KEY="pr_live_xxxxxxxxxxxxx"
 
# Launch an A100 with PyTorch
curl -X POST https://api.pulserun.dev/v1/instances \
  -H "Authorization: Bearer $PULSERUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "gpu": "a100_80gb",
    "template": "pytorch",
    "region": "us",
    "spot": false,
    "provider": "auto"
  }'

Response:

{
  "id": "inst_abc123",
  "status": "provisioning",
  "gpu": "a100_80gb",
  "gpu_name": "NVIDIA A100 80GB",
  "cost_per_hour": 0.95,
  "template": "pytorch"
}

Using the Python SDK

import pulserun
 
client = pulserun.Client("pr_live_xxxxxxxxxxxxx")
 
# Launch cheapest A100
instance = client.instances.create(
    gpu="a100_80gb",
    template="pytorch"
)
 
print(f"SSH: {instance.ssh_command}")
print(f"Jupyter: {instance.jupyter_url}")
 
# When done
instance.terminate()

What’s Next?

IntroductionAccount Setup