@handit.ai/ai-wrapper/QUICK_START.md

🤖 Intelligent AI execution system with built-in tracking, evaluation, and self-improvement capabilities. The complete AI intelligence platform for enterprise applications.

# 🚀 Quick Start Guide

## Installation

```bash
npm install @handit.ai/ai-wrapper
npm install @handit.ai/node openai @google/generative-ai @anthropic-ai/sdk
```

## Basic Setup

```javascript
const { AIWrapper } = require('@handit.ai/ai-wrapper');

const aiWrapper = new AIWrapper({
  handitApiKey: 'your-handit-api-key',
  openaiApiKey: 'your-openai-api-key',    // optional
  googleApiKey: 'your-google-api-key',    // optional
  anthropicApiKey: 'your-anthropic-key'   // optional
});
```

## 🎯 Key Methods (Choose Your Style)

### 1. `runSmartAgent()` - **Recommended for most use cases**
**The flagship method** - Full AI intelligence with tracking and self-improvement.

```javascript
const result = await aiWrapper.runSmartAgent({
  agentName: 'customer-support',
  input: 'How can I return my order?',
  model: 'gpt-4'
});

console.log(result.output);
// ✅ Tracks everything automatically
// ✅ Learns from this interaction  
// ✅ Improves future responses
```

### 2. `runOptimizedAgent()` - **For performance-focused tasks**
Google AI with optimization focus.

```javascript
const result = await aiWrapper.runOptimizedAgent({
  agentName: 'content-writer',
  input: 'Write an SEO blog post about renewable energy'
});
// ✅ Optimized for quality and performance
```

### 3. `runTrackedAgent()` - **For sensitive/compliance tasks**
Anthropic Claude with enhanced safety tracking.

```javascript
const result = await aiWrapper.runTrackedAgent({
  agentName: 'legal-reviewer',
  input: 'Review this contract for compliance issues'
});
// ✅ Enhanced safety and compliance tracking
```

### 4. Provider-Specific Methods
Direct control over AI provider.

```javascript
// OpenAI
await aiWrapper.runWithOpenAI({
  agentName: 'assistant',
  input: 'Generate code',
  model: 'gpt-4'
});

// Google AI  
await aiWrapper.runWithGoogle({
  agentName: 'analyzer',
  input: 'Analyze data',
  model: 'gemini-pro'
});

// Anthropic Claude
await aiWrapper.runWithAnthropic({
  agentName: 'reviewer',
  input: 'Review content',
  model: 'claude-3-sonnet-20240229'
});
```

## 📊 What Each Method Does

| Method | Provider | Focus | Best For |
|--------|----------|-------|----------|
| `runSmartAgent()` | OpenAI | **Intelligence** | General AI tasks, customer support, analysis |
| `runOptimizedAgent()` | Google AI | **Performance** | Content generation, optimization tasks |
| `runTrackedAgent()` | Anthropic | **Safety** | Code review, legal, compliance |
| `runWithX()` | Specific | **Control** | When you need a specific provider |

## 🎨 Quick Examples

### Basic Customer Support
```javascript
const support = await aiWrapper.runSmartAgent({
  agentName: 'customer-support',
  input: 'My order is late, what should I do?'
});
```

### Content Generation
```javascript
const content = await aiWrapper.runOptimizedAgent({
  agentName: 'blog-writer',
  input: 'Write about the future of AI in healthcare'
});
```

### Code Review
```javascript
const review = await aiWrapper.runTrackedAgent({
  agentName: 'code-reviewer',
  input: 'function login(user, pass) { /* code here */ }'
});
```

## 🔧 Environment Variables (Optional)

```bash
HANDIT_API_KEY=your-handit-key
OPENAI_API_KEY=your-openai-key
GOOGLE_API_KEY=your-google-key
ANTHROPIC_API_KEY=your-anthropic-key
```

## 🎯 Result Structure

All methods return:
```javascript
{
  success: true,                    // boolean
  output: "AI response here",       // string
  executionId: "unique-id",         // string
  prompts: [...],                   // array of prompts used
  error: null                       // string if failed
}
```

## 🚀 Next Steps

1. **Start with `runSmartAgent()`** - it's the most intelligent option
2. **Set up your agents in handit** - create prompts for your use cases
3. **Check the analytics** - see how your AI is performing
4. **Read the [full documentation](./DOCUMENTATION.md)** for advanced features

## 💡 Pro Tips

- Use descriptive `agentName` values like `'customer-support-billing'`
- The system learns from every interaction automatically
- Check the handit dashboard to see performance analytics
- All executions are tracked - no additional setup needed!

---

**Need more details?** See the [Complete Documentation](./DOCUMENTATION.md) for advanced features, best practices, and in-depth examples.