Cost Optimization

Cost Optimization

UniCraft helps you minimize AI costs while maintaining quality through intelligent cost optimization features.

Cost Management Features

Dynamic Pricing

• Real-time price monitoring across all providers
• Automatic switching to cheaper alternatives
• Price alerts and notifications

Usage Analytics

• Detailed cost breakdown by provider and model
• Usage patterns and trends analysis
• Cost forecasting and budgeting

Budget Controls

• Set spending limits per project or organization
• Automatic throttling when approaching limits
• Cost alerts and notifications

Optimization Strategies

Model Selection

Choose the most cost-effective model for your use case:

const response = await unicraft.chat.completions.create({
  messages: [{ role: "user", content: "Hello" }],
  model: "auto", // Automatically selects cost-optimized model
  max_cost: 0.01, // Maximum cost per request
  quality_threshold: 0.8, // Minimum quality requirement
});

Batch Processing

Process multiple requests together to reduce costs:

const batch = await unicraft.batch.create({
  requests: [
    { messages: [{ role: "user", content: "Request 1" }] },
    { messages: [{ role: "user", content: "Request 2" }] },
  ],
  model: "auto",
});

Caching

Implement intelligent caching to reduce redundant API calls:

const response = await unicraft.chat.completions.create({
  messages: [{ role: "user", content: "Hello" }],
  cache: true,
  cache_ttl: 3600, // Cache for 1 hour
});

Cost Monitoring

Dashboard

Monitor your spending in real-time through the UniCraft dashboard:

• Daily, weekly, and monthly cost trends
• Provider comparison charts
• Cost per request analysis

Alerts

Set up cost alerts to stay within budget:

• Daily spending limits
• Unusual usage pattern detection
• Provider price change notifications

Reports

Generate detailed cost reports:

• Export data to CSV/JSON
• Custom date ranges
• Provider and model breakdowns

Best Practices

1. Use Auto Model Selection: Let UniCraft choose the most cost-effective model
2. Implement Caching: Cache responses for frequently asked questions
3. Batch Similar Requests: Group similar requests together
4. Monitor Usage: Regularly review usage patterns and costs
5. Set Budget Limits: Use spending controls to prevent overspending