Multiyo.ai 🤖

Multiyo is a powerful, provider-agnostic AI SDK for Node.js.
It helps developers integrate AI services (like Content Moderation, Summarization, and Sentiment Analysis) without worrying about Prompt Engineering, messy API responses, or vendor lock-in.

Easily switch between Google Gemini and OpenAI with a single line of configuration.

✨ Key Features

• Unified API: Write code once, run it with Google or OpenAI.
• Strict JSON Outputs: No more Regex parsing! All services return clean, typed JSON objects.
• Pre-engineered Prompts: Optimized system instructions for common tasks.
• Lightweight: Uses Dynamic Imports to keep your project size small (loads SDKs only when needed).

📦 Installation

Install the package via npm:

npm i @elmoziml/multiyo

🔔 Note:

• The Google Gemini SDK (@google/generative-ai) is installed automatically as a dependency.
• If you plan to use OpenAI, you must install it manually:

  npm install openai

⚙️ Configuration

Multiyo works out-of-the-box with Google Gemini as the default provider. You don't need complex setup files; just use Environment Variables.

1. Set up Environment Variables

Create a .env file in your project root:

# Required for default usage
GEMINI_API_KEY=your_google_api_key_here

# Required only if you use OpenAI
OPENAI_API_KEY=your_openai_api_key_here

2. Basic Usage (Default: Google Gemini)

Initialize the service without arguments to use the default provider.

import Multiyo from '@elmoziml/multiyo';
import dotenv from 'dotenv';
dotenv.config();

const service = new Multiyo.Moderation(); 
// 🚀 Automatically connects to Google Gemini

3. Switching Providers (OpenAI)

To use a different provider, pass a configuration object:

const service = new Multiyo.Moderation({
  provider: 'openai'
});

🚀 Usage Examples

🛡️ Content Moderation

Check if a user comment is safe, toxic, or spam.

import Multiyo from '@elmoziml/multiyo';
async function checkComment() {
  const mod = new Multiyo.Moderation();
  
  try {
    const result = await mod.reviewComment("You are stupid and I hate you!");
    
    if (!result.isSafe) {
      console.log(`❌ Comment rejected: ${result.reason}`);
      console.log(`Category: ${result.category}`);
    } else {
      console.log("✅ Comment published.");
    }
  } catch (error) {
    console.error("AI Error:", error.message);
  }
}

checkComment();

Output Example:

{
  "isSafe": false,
  "reason": "Contains personal insults.",
  "category": "harassment"
}

📊 Text Analysis (Summarization & Sentiment)

Summarize long texts or detect emotions.

import Multiyo from '@elmoziml/multiyo';

async function analyzePost() {
  const analyzer = new Multiyo.Analysis(); // Default: Google

  const text = "I absolutely love this product! It changed my life completely, although it's a bit expensive.";

  // 1. Get Sentiment
  const sentiment = await analyzer.getSentiment(text);
  console.log("Sentiment:", sentiment); 
  // Output: { sentiment: "positive", score: 0.8 }

  // 2. Summarize
  const summary = await analyzer.summarize(text);
  console.log("Summary:", summary.summary);
}

📚 API Reference

ModerationService

| Method | Input | Returns | Description | | :--- | :--- | :--- | :--- | | reviewComment(text) | string | Promise<{ isSafe, reason, category }> | Analyzes text for toxicity/spam. |

AnalysisService

| Method | Input | Returns | Description | | :--- | :--- | :--- | :--- | | summarize(text) | string | Promise<{ summary, keyPoints[] }> | Summarizes text and extracts bullets. | | getSentiment(text) | string | Promise<{ sentiment, score }> | Detects positive/negative tone. |

🛠 Advanced Configuration

You can override specific settings (like the Model ID) for a specific instance.

const customService = new Multiyo.Analysis({
  provider: 'google',
  providers: {
    google: {
      model: 'gemini-1.5-pro' // Force specific model
    }
  }
});

📄 License

ISC