EasyLLM

A unified TypeScript/JavaScript SDK for OpenAI-compatible LLM APIs, including llm.vin and OpenAI.

Features

• 🚀 Universal API: Works with llm.vin, OpenAI, and any OpenAI-compatible API
• 📝 TypeScript Support: Full type safety and IntelliSense
• 🔄 Automatic Retries: Built-in retry logic for robust API calls
• 🎯 OpenAI Compatible: Drop-in replacement for OpenAI SDK syntax
• 🛡️ Error Handling: Comprehensive error handling and validation
• 🔍 Content Moderation: Built-in support for text and image moderation
• 🌐 Web Search: Built-in web search capabilities for real-time information
• ⚡ Streaming Support: Real-time streaming responses with function calling

Installation

npm install @llmvin/easyllm

Quick Start

import EasyLLM from '@llmvin/easyllm';

const client = new EasyLLM({
  apiKey: 'your-api-key'
});

const response = await client.chat.completions.create({
  model: 'llama4-scout',
  messages: [
    { role: 'user', content: 'Hello, world!' }
  ]
});

console.log(response.choices[0].message.content);

Web Search

EasyLLM includes built-in web search capabilities that allow AI models to access real-time information from the web.

Note: Web search is currently supported by the llama4-maverick model on llm.vin.

Basic Web Search

import EasyLLM from '@llmvin/easyllm';

const client = new EasyLLM({
  apiKey: 'your-api-key',
  webSearch: {
    enabled: true,
    maxResults: 5,
    includeContent: true
  }
});

// Automatic web search when enabled globally
const response = await client.chat.completions.createWithWebSearch({
  model: 'llama4-maverick',
  messages: [
    { role: 'user', content: 'What are the latest developments in AI?' }
  ]
});

// Or enable web search for specific requests
const response2 = await client.chat.completions.create({
  model: 'llama4-maverick',
  messages: [
    { role: 'user', content: 'What is the current weather in Tokyo?' }
  ],
  webSearch: true // Enable for this request only
});

Streaming with Web Search

const stream = await client.chat.completions.streamWithWebSearch({
  model: 'llama4-maverick',
  messages: [
    { role: 'user', content: 'Tell me about recent space missions' }
  ]
});

for await (const chunk of stream) {
  if (chunk.choices[0]?.delta?.content) {
    process.stdout.write(chunk.choices[0].delta.content);
  }
}

Dynamic Web Search Control

// Enable/disable web search at runtime
client.setWebSearchEnabled(true);
console.log(client.isWebSearchEnabled()); // true

client.setWebSearchEnabled(false);
console.log(client.isWebSearchEnabled()); // false

Supported APIs

• llm.vin - Access to multiple AI models (default)
• OpenAI - Official OpenAI API
• Custom - Any OpenAI-compatible endpoint

Documentation

• Complete API Documentation
• Usage Examples

⚠️ AI Development Notice

This project was developed with significant AI assistance:

• 🤖 Code Review: Source code was read, analyzed, and improved by AI
• 📝 Documentation: All documentation (README, examples, API docs) was written by AI
• 💬 Commit Messages: Git commit messages were generated by AI
• 🧪 Tests: Test suites were designed and implemented with AI assistance
• 🏗️ Architecture: Project structure and design patterns suggested by AI

Please be aware:

• While the code has been tested, AI-generated code should be thoroughly reviewed before production use
• Documentation may contain inaccuracies or assumptions that need verification
• Always validate functionality against your specific use cases
• Consider this a starting point that may require human review and refinement

This notice is provided for transparency about the development process.

License

MIT © llm.vin