AI-Powered Cost Insights

Get intelligent recommendations and insights to optimize your cloud spending with our privacy-first AI assistant.

100% Privacy-First AI

Your cost data never leaves your browser. All AI processing happens locally on your device using WebGPU technology.

  • No data sent to external AI services
  • Powered by Llama 3.2 3B running in your browser
  • Complete control over your sensitive cost information

How It Works

1. WebGPU Technology

Our AI uses WebGPU, a modern browser API that provides high-performance access to your device's GPU. WebGPU is now widely supported across all modern browsers, enabling sophisticated AI models to run directly in your browser without any server-side processing.

2. Local Model Execution

The Llama 3.2 3B model (optimized for efficiency) is downloaded once and cached in your browser. All subsequent AI analysis happens locally - your cost data never leaves your machine.

3. Real-Time Insights

Click the AI sparkles icon on any dashboard section to get instant, context-aware recommendations for optimizing costs, identifying anomalies, and reducing cloud spending.

AI-Powered Features

Cost Trend Analysis

Identify spending patterns, predict future costs, and get alerts about unusual spikes before they impact your budget.

Provider Optimization

Analyze cost distribution across cloud providers and receive recommendations for better resource allocation.

Service Breakdown

Get insights into which services are consuming the most budget and discover opportunities for right-sizing or elimination.

Activity Monitoring

Understand spending changes related to deployments, migrations, and infrastructure updates with AI-powered correlation analysis.

System Requirements

AI features work on all modern browsers with WebGPU support:

  • Chrome/Edge 113+
  • Firefox 121+ (WebGPU enabled by default)
  • Safari 18+ on macOS Sonoma or later

The first time you use AI features, the model (~1.8GB) will be downloaded and cached. Subsequent uses will be instant.

Managing AI Settings

You have complete control over AI features. To enable or disable AI suggestions:

  1. 1.Navigate to Settings from your dashboard
  2. 2.Click the Preferences tab
  3. 3.Scroll to the AI Features section
  4. 4.Toggle Enable AI Suggestions on or off

When disabled, all AI suggestion buttons will be hidden throughout your dashboard. This setting is personal to your user account and does not affect other team members.

Frequently Asked Questions

Is my cost data sent to any external servers?

No, absolutely not. All AI processing happens locally in your browser using WebGPU. Your cost data never leaves your device.

Why is the first AI request slow?

The first time you use AI features, the Llama 3.2 3B model (~1.8GB) needs to be downloaded and initialized. This happens once, and the model is cached in your browser. All subsequent requests are nearly instant.

What if my browser doesn't support WebGPU?

WebGPU is now widely supported across all modern browsers. If you encounter any issues, make sure your browser is up to date. The rest of the CosmosCost platform works perfectly without AI features - they're an optional enhancement.

Does AI use affect my device performance?

AI processing uses your GPU efficiently and only runs when you explicitly request insights. When not in use, it has zero impact on your device performance or battery life.