GPU Arbitrage & Availability Scanner

Find the cheapest GPU compute across multiple cloud providers in seconds — designed for AI agents and developers.

Stop wasting time manually comparing GPU prices across different cloud platforms. This Actor automatically scans RunPod, Lambda Labs, Vast.ai, AWS EC2, and Google Cloud Platform to find you the best deals on H100s, A100s, and other AI-optimized GPUs.

🚀 What You Get

• Real-time price comparison across 5 major GPU cloud providers
• AI-optimized output with value scoring for instant decision making
• Live availability checking - only see GPUs you can actually book
• Smart filtering by GPU model, max price, and provider preference
• MCP integration - works seamlessly with Claude, GPT, and other AI assistants

💡 Use Cases

• AI Researchers: Quickly find cheapest H100 clusters for training runs
• Startups: Optimize GPU costs while scaling ML workloads
• Freelancers: Compare spot prices before spinning up inference servers
• AI Agents: Automated GPU procurement for autonomous systems
• Cost Optimization: Track pricing trends across providers

🎯 Quick Start

Via Apify Console

1. Click "Start" with default settings
2. Results appear in your dataset within 30 seconds
3. Export to JSON, CSV, or connect to your app via API

Via AI Assistant (MCP)

Simply ask your AI assistant:

"Find me the cheapest H100 GPU under $3/hour"

The assistant will automatically use this Actor and present formatted results.

Via API

curl -X POST "https://api.apify.com/v2/acts/YOUR_USERNAME~gpu-arbitrage-mcp/runs" \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "gpuFilter": ["H100", "A100"],
    "maxPrice": 4.0,
    "providers": ["RunPod", "Vast", "AWS"]
  }'

🔧 Input Options

Parameter	Type	Default	Description
gpuFilter	array	[]	GPU models to find (e.g., ["H100", "A100"])
maxPrice	number	10.0	Maximum price per hour in USD
providers	array	All providers	Choose from: RunPod, Lambda, Vast, AWS, GCP
onlyAvailable	boolean	true	Only show immediately bookable GPUs

📊 Output Format

Get clean, structured data sorted by best value:

{
  "provider": "RunPod",
  "gpuModel": "H100",
  "vram": 80,
  "price": 2.49,
  "priceType": "spot",
  "region": "US East",
  "available": true,
  "valueScore": 0.0311,
  "link": "https://runpod.io/console/deploy?..."
}

🎨 Smart Value Scoring

The valueScore field helps you compare different GPU configurations:

• H100 80GB @ $4/hr → valueScore: 0.050
• A100 40GB @ $1.50/hr → valueScore: 0.038 ← Better value!

Lower scores = better price-to-performance ratio.

🔌 Integrations

AI Assistants

Works out-of-the-box with:

• Claude (via MCP server)
• GPT-4 (via API)
• Custom agents (via webhook)

Development

// Get data in your app
const response = await fetch('https://api.apify.com/v2/datasets/YOUR_DATASET_ID/items');
const gpuDeals = await response.json();

// Find best deal
const bestGpu = gpuDeals[0]; // Already sorted by value
console.log(`Best deal: ${bestGpu.gpuModel} @ $${bestGpu.price}/hr`);

🌍 Supported Providers

• RunPod - Real-time spot and on-demand pricing
• Lambda Labs - GPU cloud instances
• Vast.ai - Community GPU marketplace
• AWS EC2 - P4d, P3, G4dn, G5 instances
• Google Cloud Platform - Compute Engine GPU instances

🚦 Current Status

✅ Real-time data from RunPod and Vast.ai ✅ Live availability checking ✅ 5 provider coverage ✅ Static pricing for AWS and GCP (accurate as of Dec 2024) ✅ Production ready for immediate use

💰 Pricing

This Actor runs on Apify's platform pricing:

• Free tier: 1,000 Actor hours/month
• Pay as you go: $0.25 per hour of compute
• Typical run: Completes in 5-30 seconds (~$0.002-0.01 per search)

🆘 Support

• Documentation: Full guides in your Apify Console
• Issues: Report bugs via the Actor's issue tracker
• Feature requests: Contact via Apify platform

⚡ Pro Tips

1. Use GPU filters to avoid irrelevant results: {"gpuFilter": ["H100"]}
2. Set realistic price limits to focus on bookable options
3. Check multiple providers - prices vary significantly
4. Book quickly - spot prices can change within minutes
5. Integrate with alerts to get notified of good deals

---

© 2024 GPU Arbitrage Scanner | Professional GPU Price Intelligence