Multi-Provider

This guide demonstrates how to use multiple AI providers with Robota, showcasing different models and comparing their responses.

Overview

The multi-provider example shows how to:

• Configure multiple OpenAI models (GPT-3.5, GPT-4o-mini)
• Create independent agent instances with different providers
• Compare responses from different models
• Monitor individual agent statistics

Code Example

/**
 * 03-multi-providers.ts
 *
 * This example demonstrates using multiple AI providers:
 * - OpenAI with different models
 * - Comparing responses from different models
 * - Provider switching capabilities
 * - Independent agent instances
 */

import OpenAI from 'openai';
import { Robota } from '@robota-sdk/agent-core';
import { OpenAIProvider } from '@robota-sdk/agent-provider-openai';
import dotenv from 'dotenv';

// Load environment variables from examples directory
dotenv.config();

async function testProvider(providerName: string, robota: Robota, query: string) {
  console.log(`\n${'='.repeat(50)}`);
  console.log(`🤖 Testing ${providerName.toUpperCase()} Provider`);
  console.log(`${'='.repeat(50)}`);
  console.log(`User: ${query}`);

  try {
    const response = await robota.run(query);
    console.log(`Assistant: ${response}`);
  } catch (error) {
    console.error(`❌ Error with ${providerName}:`, error);
  }
}

async function main() {
  try {
    console.log('🌐 Multi-Provider Example Started...\n');

    // Validate API key
    const openaiKey = process.env.OPENAI_API_KEY;
    if (!openaiKey) {
      throw new Error('OPENAI_API_KEY environment variable is required for this example');
    }

    // Create OpenAI client
    const openaiClient = new OpenAI({ apiKey: openaiKey });

    // === OpenAI GPT-3.5-turbo Provider Test ===
    const openai35Provider = new OpenAIProvider({
      apiKey: openaiKey,
    });

    const robota35 = new Robota({
      name: 'GPT35Agent',
      aiProviders: [openai35Provider],
      defaultModel: {
        provider: 'openai',
        model: 'gpt-3.5-turbo',
        systemMessage:
          'You are a helpful assistant powered by OpenAI GPT-3.5. Be concise and mention that you are GPT-3.5.',
      },
    });

    await testProvider(
      'OpenAI GPT-3.5',
      robota35,
      'Hello! Please tell me about artificial intelligence in 2-3 sentences.',
    );

    // === OpenAI GPT-4o-mini Provider Test ===
    const openai4MiniProvider = new OpenAIProvider({
      apiKey: openaiKey,
    });

    const robota4Mini = new Robota({
      name: 'GPT4MiniAgent',
      aiProviders: [openai4MiniProvider],
      defaultModel: {
        provider: 'openai',
        model: 'gpt-4o-mini',
        systemMessage:
          'You are a helpful assistant powered by OpenAI GPT-4o-mini. Be detailed and mention that you are GPT-4o-mini.',
      },
    });

    await testProvider(
      'OpenAI GPT-4o-mini',
      robota4Mini,
      'Hello! Please tell me about artificial intelligence in 2-3 sentences.',
    );

    // === Test different model comparison ===
    console.log(`\n${'='.repeat(50)}`);
    console.log('🔄 Testing Model Comparison');
    console.log(`${'='.repeat(50)}`);

    const testQueries = ['What is AI?'];

    for (const query of testQueries) {
      console.log(`\n📝 Query: ${query}`);

      console.log('\n🤖 GPT-3.5-turbo:');
      try {
        const response35 = await robota35.run(query);
        console.log(response35);
      } catch (error) {
        console.error('Error:', error);
      }

      console.log('\n🧠 GPT-4o-mini:');
      try {
        const response4 = await robota4Mini.run(query);
        console.log(response4);
      } catch (error) {
        console.error('Error:', error);
      }

      console.log('\n' + '-'.repeat(40));
    }

    // === Show Agent Statistics ===
    console.log('\n📊 Agent Comparison:');

    const stats35 = robota35.getStats();
    console.log(`\nGPT-3.5 Agent (${stats35.name}):`);
    console.log(`- History length: ${stats35.historyLength}`);
    console.log(`- Uptime: ${Math.round(stats35.uptime)}ms`);

    const stats4 = robota4Mini.getStats();
    console.log(`\nGPT-4o-mini Agent (${stats4.name}):`);
    console.log(`- History length: ${stats4.historyLength}`);
    console.log(`- Uptime: ${Math.round(stats4.uptime)}ms`);

    console.log('\n✅ Multi-Provider Example Completed!');
    console.log('\n💡 To test other providers (Anthropic, Google AI):');
    console.log('   - Set ANTHROPIC_API_KEY environment variable');
    console.log('   - Set GOOGLE_AI_API_KEY environment variable');
    console.log(
      '   - Import @robota-sdk/agent-provider-anthropic and @robota-sdk/agent-provider-google packages',
    );

    // Clean up resources
    await robota35.destroy();
    await robota4Mini.destroy();

    console.log('🧹 Cleanup completed. Exiting...');
    process.exit(0);
  } catch (error) {
    console.error('❌ Error occurred:', error);
    process.exit(1);
  }
}

// Execute
main();

Expected Output

🌐 Multi-Provider Example Started...

==================================================
🤖 Testing OPENAI GPT-3.5 Provider
==================================================
User: Hello! Please tell me about artificial intelligence in 2-3 sentences.
Assistant: Hello! I'm GPT-3.5, an AI assistant. Artificial intelligence (AI) refers to computer systems that can perform tasks typically requiring human intelligence, such as learning, reasoning, and problem-solving. AI technologies power everything from voice assistants to autonomous vehicles, and continue to advance rapidly across various industries.

==================================================
🤖 Testing OPENAI GPT-4O-MINI Provider
==================================================
User: Hello! Please tell me about artificial intelligence in 2-3 sentences.
Assistant: Hello! I'm GPT-4o-mini, an advanced AI assistant. Artificial intelligence is a field of computer science focused on creating systems that can perform tasks requiring human-like intelligence, including learning from data, recognizing patterns, and making decisions. AI has revolutionized numerous industries by enabling automation, predictive analytics, and intelligent human-computer interactions, from chatbots and recommendation systems to medical diagnosis and autonomous vehicles.

==================================================
🔄 Testing Model Comparison
==================================================

📝 Query: What is AI?

🤖 GPT-3.5-turbo:
AI, or Artificial Intelligence, refers to computer systems designed to perform tasks that typically require human intelligence...

🧠 GPT-4o-mini:
AI (Artificial Intelligence) is a comprehensive field of computer science that focuses on creating intelligent machines...

📊 Agent Comparison:

GPT-3.5 Agent (GPT35Agent):
- History length: 4
- Uptime: 2156ms

GPT-4o-mini Agent (GPT4MiniAgent):
- History length: 4
- Uptime: 2287ms

✅ Multi-Provider Example Completed!

Key Features

1. Independent Agent Instances

Each provider has its own Robota instance with separate configuration:

// GPT-3.5 instance
const robota35 = new Robota({
  name: 'GPT35Agent',
  aiProviders: [openai35Provider],
  defaultModel: {
    provider: 'openai',
    model: 'gpt-3.5-turbo',
    systemMessage: 'You are a helpful assistant powered by OpenAI GPT-3.5.',
  },
});

// GPT-4o-mini instance
const robota4Mini = new Robota({
  name: 'GPT4MiniAgent',
  aiProviders: [openai4MiniProvider],
  defaultModel: {
    provider: 'openai',
    model: 'gpt-4o-mini',
    systemMessage: 'You are a helpful assistant powered by OpenAI GPT-4o-mini.',
  },
});

2. Provider Testing Helper

Reusable function to test different providers:

async function testProvider(providerName: string, robota: Robota, query: string) {
  console.log(`🤖 Testing ${providerName.toUpperCase()} Provider`);
  console.log(`User: ${query}`);

  try {
    const response = await robota.run(query);
    console.log(`Assistant: ${response}`);
  } catch (error) {
    console.error(`❌ Error with ${providerName}:`, error);
  }
}

3. Model Comparison

Side-by-side comparison of model responses:

console.log('\n🤖 GPT-3.5-turbo:');
const response35 = await robota35.run(query);
console.log(response35);

console.log('\n🧠 GPT-4o-mini:');
const response4 = await robota4Mini.run(query);
console.log(response4);

4. Statistics Comparison

Compare performance metrics across different agents:

const stats35 = robota35.getStats();
const stats4 = robota4Mini.getStats();

console.log(`GPT-3.5 Agent: ${stats35.historyLength} messages, ${stats35.uptime}ms uptime`);
console.log(`GPT-4o-mini Agent: ${stats4.historyLength} messages, ${stats4.uptime}ms uptime`);

Advanced Multi-Provider Setup

Adding Anthropic Provider

import { AnthropicProvider } from '@robota-sdk/agent-provider-anthropic';

const anthropicProvider = new AnthropicProvider({
  apiKey: process.env.ANTHROPIC_API_KEY!,
});

const robotaClaude = new Robota({
  name: 'ClaudeAgent',
  aiProviders: [anthropicProvider],
  defaultModel: {
    provider: 'anthropic',
    model: 'claude-3-haiku-20240307',
    systemMessage: 'You are Claude, an AI assistant by Anthropic.',
  },
});

Adding Google AI Provider

import { GoogleProvider } from '@robota-sdk/agent-provider-google';

const googleProvider = new GoogleProvider({
  apiKey: process.env.GOOGLE_AI_API_KEY!,
});

const robotaGemini = new Robota({
  name: 'GeminiAgent',
  aiProviders: [googleProvider],
  defaultModel: {
    provider: 'google',
    model: 'gemini-1.5-flash',
    systemMessage: 'You are Gemini, an AI model by Google.',
  },
});

Best Practices

1. Environment Variables: Use environment variables for API keys
2. Error Handling: Implement proper error handling for each provider
3. Resource Management: Always call destroy() on each agent instance
4. Provider Naming: Use clear, descriptive names for different provider instances
5. Statistics Monitoring: Track performance metrics for comparison

Provider Support

Currently supported providers:

• OpenAI: GPT-3.5, GPT-4, GPT-4o-mini and other models
• Anthropic: Claude 3 family (Haiku, Sonnet, Opus)
• Google AI: Gemini 1.5 (Flash, Pro) models

Each provider maintains feature parity through the unified BaseAIProvider architecture.