AI Service SDK

🚀 A unified Node.js SDK for integrating multiple AI service providers with a consistent API interface.

여러 AI 서비스 제공업체를 일관된 API 인터페이스로 통합하는 통합 Node.js SDK입니다.

✨ Key Features

• 🤖 Multi-provider support: OpenAI, Anthropic, Google Gemini, Stability AI, Runway, ElevenLabs
• 🔄 Unified API: Consistent interface across all providers
• 💬 Advanced Conversation Management: Support for both conversation history and server-side conversation state
• 🌐 Web Search Integration: Real-time web search capabilities with integrated AI processing
• 🛠️ Custom Tools: Register and use custom functions with AI models
• 💰 Cost Tracking: Automatic cost calculation for API usage (USD & KRW)
• ⚙️ Environment-based Configuration: Easy setup with environment variables
• 🧪 Comprehensive Testing: Built-in test suite for all features

🎯 Supported AI Providers

| Provider | Text | Image | TTS | Video | Web Search | Conversation | |----------|------|-------|-----|-------|------------|-------------| | OpenAI | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ (History + State) | | Anthropic | ✅ | ❌ | ❌ | ❌ | ✅ | ✅ (History) | | Google Gemini | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ (History) | | Stability AI | ❌ | ✅ | ❌ | ❌ | ❌ | ❌ | | Runway | ❌ | ❌ | ❌ | ✅ | ❌ | ❌ | | ElevenLabs | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ |

🔧 Installation

pnpm add @librorum/aiservice
# or
npm install @librorum/aiservice
# or
yarn add @librorum/aiservice

🌍 Environment Setup

Create a .env file in your project root:

OPENAI_API_KEY=your_openai_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key
GEMINI_API_KEY=your_gemini_api_key
STABILITY_API_KEY=your_stability_api_key
RUNWAY_API_KEY=your_runway_api_key
ELEVENLABS_API_KEY=your_elevenlabs_api_key

🚀 Quick Start

import aiService from '@librorum/aiservice';

// Check API key status
const api_status = aiService.testApiKeys();
console.log(api_status);

// Simple text generation
const { text } = await aiService.generateText({
  provider: 'openai',
  prompt: 'Hello, world!'
});

console.log(text);

💬 Conversation Management

Method 1: Conversation History (All Providers)

// Start a conversation
let conversation_history = [];

// First message
const response1 = await aiService.generateText({
  provider: 'openai', // or 'anthropic', 'gemini'
  prompt: 'My name is John. Remember this.',
  conversation_history
});

conversation_history = response1.conversation_history;
console.log(response1.text); // "I'll remember that your name is John."

// Continue conversation
const response2 = await aiService.generateText({
  provider: 'openai',
  prompt: 'What is my name?',
  conversation_history
});

console.log(response2.text); // "Your name is John."

Method 2: Conversation State (OpenAI Only)

// First message with conversation state
const response1 = await aiService.generateText({
  provider: 'openai',
  prompt: 'My name is John. Remember this.',
  use_conversation_state: true,
});

console.log(response1.response_id); // "resp_xxx..."

// Continue with response ID
const response2 = await aiService.generateText({
  provider: 'openai',
  prompt: 'What is my name?',
  use_conversation_state: true,
  previous_response_id: response1.response_id
});

console.log(response2.text); // "Your name is John."

🎨 Text Generation

// Basic usage
const { text, usage, cost } = await aiService.generateText({
  provider: 'openai',
  prompt: 'Explain quantum computing'
});

// Advanced options
const response = await aiService.generateText({
  provider: 'anthropic',
  model: 'claude-sonnet-4-0',
  prompt: 'Write a creative story',
  temperature: 0.8,
  max_tokens: 1000,
  ai_rule: 'Write in a friendly tone'
});

// With web search
const { text: searchResult } = await aiService.generateText({
  provider: 'gemini',
  prompt: 'Find recent AI news and summarize',
  system_tools: ['web_search']
});

🖼️ Image Generation

import fs from 'fs/promises';

// Basic image generation
const { image } = await aiService.generateImage({
  provider: 'openai',
  prompt: 'A futuristic city with flying cars'
});

await fs.writeFile('image.jpg', image);

// Custom parameters
const { image: customImage } = await aiService.generateImage({
  provider: 'stability',
  model: 'stable-diffusion-v1-5',
  prompt: 'A serene mountain landscape',
  width: 1024,
  height: 1024
});

🔊 Text-to-Speech

import fs from 'fs/promises';

// OpenAI TTS
const { audio } = await aiService.generateTTS({
  provider: 'openai',
  prompt: 'Hello, this is a test.',
  voice: 'alloy'
});

await fs.writeFile('speech.mp3', audio);

// ElevenLabs TTS
const { audio: elevenAudio } = await aiService.generateTTS({
  provider: 'elevenlabs',
  prompt: 'Hello, this is ElevenLabs.',
  voice_id: 'your_voice_id'
});

🎬 Video Generation

// Runway video generation
const { video } = await aiService.generateVideo({
  provider: 'runway',
  input: {
    prompt: 'A serene lake with mountains',
    duration: 5
  }
});

await fs.writeFile('video.mp4', video);

🛠️ Custom Tools

// Register a calculator tool
aiService.registerTool({
  name: 'calculator',
  description: 'Evaluates mathematical expressions',
  parameters: {
    type: 'object',
    properties: {
      expression: {
        type: 'string',
        description: 'Mathematical expression to evaluate'
      }
    },
    required: ['expression']
  },
  func: async ({ expression }) => {
    try {
      return Function(`"use strict"; return (${expression})`)()
    } catch (error) {
      return "Invalid expression"
    }
  }
});

// Use the tool
const { text } = await aiService.generateText({
  provider: 'openai',
  prompt: 'Calculate 15 * 23 + 45',
  user_tools: ['calculator']
});

🧪 Testing

// Test all providers
await aiService.test();

// Test specific provider
await aiService.test('openai');

// Test specific feature
await aiService.test('gemini', 'text');

// Test with tools
await aiService.test('openai', 'text', ['web_search'], ['calculator']);

NPM Test Scripts

# Basic tests
pnpm test                        # Test all features
pnpm run test:openai            # Test OpenAI
pnpm run test:websearch         # Test web search

# Conversation tests
pnpm run test:conversation       # Test all providers
pnpm run test:conversation:openai        # OpenAI history
pnpm run test:conversation:anthropic     # Anthropic history
pnpm run test:conversation:gemini        # Gemini history
pnpm run test:conversation_state         # OpenAI state mode

# Media tests
pnpm run test:stability         # Image generation
pnpm run test:elevenlabs        # Text-to-speech
pnpm run test:runway           # Video generation

💰 Cost Tracking

const { text, usage, cost } = await aiService.generateText({
  provider: 'openai',
  prompt: 'Hello world'
});

console.log(usage); // { input_tokens: 10, output_tokens: 5, total_tokens: 15 }
console.log(cost);  // { usd: 0.0001, krw: 0.13 }

API Reference

generateText(options)

{{ ... }} Generate text using various AI providers.

다양한 AI 제공업체를 사용하여 텍스트를 생성합니다.

Parameters:

• provider (string): AI provider ('openai', 'anthropic', 'gemini')
• model (string): Model name
• prompt (string): Input prompt
• ai_rule (string, optional): Additional AI instructions
• temperature (number, optional): Creativity level (0-1)
• max_tokens (number, optional): Maximum response length
• web_search (boolean, optional): Enable web search integration
• system_tools (array, optional): System tools to use (e.g., ['web_search'])
• user_tools (array, optional): User tools to use (e.g., ['calculator'])

generateImage(options)

Generate images using AI providers.

AI 제공업체를 사용하여 이미지를 생성합니다.

Parameters:

• provider (string): AI provider ('openai', 'stability')
• model (string): Model name
• prompt (string): Image description
• width (number, optional): Image width
• height (number, optional): Image height
• n (number, optional): Number of images to generate

generateTTS(options)

Convert text to speech.

텍스트를 음성으로 변환합니다.

Parameters:

• provider (string): AI provider ('openai', 'elevenlabs')
• model (string): Model name
• prompt (string): Text to convert
• voice (string, optional): Voice type (OpenAI)
• voice_id (string, optional): Voice ID (ElevenLabs)
• response_format (string, optional): Audio format

generateVideo(options)

Generate videos using AI providers.

AI 제공업체를 사용하여 비디오를 생성합니다.

Parameters:

• provider (string): AI provider ('runway')
• model (string): Model name
• input (object): Video generation parameters

test(providerName, feature, system_tools, user_tools)

Test AI provider functionality with specific features and tools.

특정 기능과 도구로 AI 제공업체 기능을 테스트합니다.

Parameters:

• providerName (string, optional): Provider name to test. If null, tests all providers
• feature (string, optional): Feature to test ('text', 'image', 'audio', 'video'). If null, tests all features
• system_tools (array, optional): System tools to include in test (e.g., ['web_search'])
• user_tools (array, optional): User tools to include in test (e.g., ['calculator'])

Returns:

• Promise that resolves to test results object

API Pricing Information

• Google Gemini: https://ai.google.dev/gemini-api/docs/pricing
• Anthropic: https://www.anthropic.com/pricing#api
• OpenAI: https://openai.com/api/pricing/

API 가격 정보

License

MIT

라이선스: MIT

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

기여를 환영합니다! 언제든지 Pull Request를 제출해 주세요.