README

Overview

The Chat AI Laravel Package is an intelligent chatbot integration for Laravel applications. It retrieves responses dynamically from multiple sources, ensuring accuracy and efficiency:

1. Database – Checks for an existing response.
2. ChatGPT – If no response is found in the database, it queries OpenAI’s ChatGPT.
3. Gemini AI – If ChatGPT reaches its limit, it switches to Google Gemini.
4. Google Search – If no AI-generated response is available, it fetches relevant search results.
5. Self-Learning – If an answer is not found, it is stored in the database for future queries.

How It Works

The chatbot follows a structured search approach:

1. Database Lookup

   • The chatbot first searches the database for a response.
   • Admins can manually add a list of questions and answers.
   • If the user's question is at least 70% similar (adjustable via the config file) to a stored question, it retrieves the answer.

2. ChatGPT Integration

   • If no answer is found in the database, it queries ChatGPT.
   • The response is then stored in the database to reduce API usage and allow admin modifications.

3. Fallback to Gemini AI

   • If ChatGPT is inactive or has reached its limit, the chatbot switches to Gemini AI and follows the same process.

4. Google Search

   • If neither ChatGPT nor Gemini AI provides an answer, it fetches relevant data using Google Search.

Provides

• Simple Blade View – A minimal and customizable chat interface.
• Lightweight Controller – A clean structure to handle requests and responses.
• Database Migration – Stores previous queries and responses for improved efficiency.

Features

• Multi-source response system (Database → ChatGPT → Gemini → Google Search).
• Self-learning AI – Stores new questions and answers.
• Efficient fallback mechanism for uninterrupted responses.
• Simple API interface for easy integration.
• Optimized for performance with caching and similarity checks.

Installation

1. Install via Composer

Run the following command to install the package:

composer require mohammadsalamat/chat_ai

2. Publish Configuration (Optional)

php artisan vendor:publish

This command publishes the migration and view files.

3. Run Migrations

php artisan migrate

This creates a new database table called questions to store chatbot interactions.

4. Add API Keys to .env

Obtain API keys for OpenAI, Gemini, and Google Search, then update your .env file:

CHATGPT_API_KEY=your_chatgpt_api_key
GEMINI_API_KEY=your_gemini_api_key
GOOGLE_SEARCH_API_KEY=your_google_search_api_key
GOOGLE_SEARCH_CX=your_google_search_cx

Usage

1. Use in a Controller or Route

Dependency Injection Approach

use Salamat\chat_ai\Http\Controllers\ChatAiController;

public function sendMessage(Request $request, ChatAiController $controller)
{
    return $controller->generateText($request);
}

Instantiating the Controller Manually

use Salamat\chat_ai\Http\Controllers\ChatAiController;

public function sendMessage(Request $request)
{
    $controller = new ChatAiController();
    return $controller->generateText($request);
}

2. Create a POST Route to Send a Message

use App\Http\Controllers\Controller;
use Illuminate\Support\Facades\Route;

Route::post('/sendMessage', [Controller::class, 'sendMessage']);

3. Frontend Integration

Send an AJAX request to your Laravel endpoint. A custom view is available at resources/views/chat_ai, including the following AJAX request:

fetch('/sendMessage', {
    method: 'POST',
    headers: {
        'Content-Type': 'application/json',
        'X-CSRF-TOKEN': '{{ csrf_token() }}' // CSRF token for security
    },
    body: JSON.stringify({
        message: messageText
    })
})
.then(response => response.json())
.then(data => console.log(data));

Configuration

The package includes a configuration file located at config/chat_ai.php, where you can customize settings such as:

• API Limits
• Similarity Threshold
• Caching Options

Contributing

Feel free to submit pull requests or open issues for bug fixes and feature requests.

License

This package is open-source and available under the MIT License.