deca-chat

DecaChat

A lightweight and easy-to-use wrapper for OpenAI's Chat API. DecaChat provides a simple interface for creating chat-based applications with OpenAI's GPT models.

Features

• 🚀 Simple, intuitive API
• 📝 TypeScript support
• 💾 Conversation management
• ⚙️ Configurable parameters
• 🛡️ Built-in error handling
• 🌐 Custom base URL support
• 🔄 Conversation history management
• 🤖 System message configuration
• 👋 Customizable introduction messages
• 📦 Zero dependencies (except OpenAI SDK)

Installation

npm install deca-chat

Quick Start

import { DecaChat } from 'deca-chat';

// Initialize the chat client
const chat = new DecaChat({
  apiKey: 'your-openai-api-key'
});

// Send a message and get a response
async function example() {
  const response = await chat.chat('Hello, how are you?');
  console.log(response);
}

Configuration

The DecaChat constructor accepts a configuration object with the following options:

interface DecaChatConfig {
  apiKey: string;         // Required: Your OpenAI API key
  model?: string;         // Optional: Default 'gpt-4o-mini'
  baseUrl?: string;       // Optional: Default 'https://api.openai.com/v1'
  maxTokens?: number;     // Optional: Default 1000
  temperature?: number;   // Optional: Default 0.7
  intro?: string;         // Optional: Custom introduction message
  systemMessage?: string; // Optional: Initial system message
  useBrowser?: boolean;   // Optional: Enable browser usage (Ensure API keys are secured!)
}

API Reference

Constructor

const chat = new DecaChat(config: DecaChatConfig);

Methods

setSystemMessage(message: string): void

Sets the system message for the conversation. This resets the conversation history and starts with the new system message.

chat.setSystemMessage('You are a helpful assistant specialized in JavaScript.');

setIntro(message: string): void

Sets a custom introduction message that will be sent to the user when starting a new conversation.

chat.setIntro('Hi! I'm your AI assistant. How can I help you today?');

async chat(message: string): Promise<string>

Sends a message and returns the assistant's response. The message and response are automatically added to the conversation history.

const response = await chat.chat('What is a closure in JavaScript?');

clearConversation(): void

Clears the entire conversation history, including the system message.

chat.clearConversation();

getConversation(): ChatMessage[]

Returns the current conversation history, including system messages, user messages, and assistant responses.

const history = chat.getConversation();

Example Usage

Basic Chat Application

import { DecaChat } from 'deca-chat';

async function example() {
  // Initialize with custom configuration including system message
  const chat = new DecaChat({
    apiKey: 'your-openai-api-key',
    model: 'gpt-4',
    maxTokens: 2000,
    temperature: 0.8,
    intro: 'Hello! I'm your coding assistant. Ask me anything about programming!',
    systemMessage: 'You are a helpful coding assistant specialized in JavaScript.'
  });

  // The system message can also be set after initialization
  chat.setSystemMessage('You are now a Python expert.');

  // The intro message will be automatically sent when starting a conversation
  const response = await chat.chat('How do I create a React component?');
  console.log('Assistant:', response);
}

Custom API Endpoint

const chat = new DecaChat({
  apiKey: 'your-api-key',
  baseUrl: 'https://your-custom-endpoint.com/v1',
  model: 'gpt-4o-mini'
});

Managing Conversations

// Start with a system message
chat.setSystemMessage('You are a helpful assistant.');

// Have a conversation
await chat.chat('What is TypeScript?');
await chat.chat('How does it differ from JavaScript?');

// Get the conversation history
const history = chat.getConversation();
console.log(history);
/* Output:
[
  { role: 'system', content: 'You are a helpful assistant.' },
  { role: 'user', content: 'What is TypeScript?' },
  { role: 'assistant', content: '...' },
  { role: 'user', content: 'How does it differ from JavaScript?' },
  { role: 'assistant', content: '...' }
]
*/

// Clear the conversation and start fresh
chat.clearConversation();

Error Handling

The chat method throws errors when:

• The API key is invalid
• The API request fails
• Rate limits are exceeded
• Network errors occur
• Invalid model specified
• Custom endpoint is unreachable

Always wrap API calls in try-catch blocks for proper error handling:

try {
  const response = await chat.chat('Hello');
  console.log(response);
} catch (error) {
  console.error('Chat error:', error);
}

Best Practices

• System Messages: Set appropriate system messages to guide the assistant's behavior
• Conversation Management: Clear conversations when starting new topics
• Error Handling: Always implement proper error handling
• Resource Management: Monitor token usage and conversation length
• API Key Security: Never expose your API key in client-side code

Contributing

Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.

License

MIT License - see the LICENSE file for details.