QuéAI - Smart LLM Chatbot

A knowledge-based chatbot designed for seamless website integration, leveraging business information from text files to provide automated customer support. This chatbot is ideal for small businesses, e-commerce platforms, and customer service teams who need an efficient way to handle inquiries 24/7. It is particularly useful for automating responses to common questions, reducing response times, and improving user experience.

Why Choose This Chatbot?

• Unique Feature: QuéAI uses Natural Language Processing to understand free-form customer queries, recognize various phrasings, slang, and context — enabling natural and seamless conversations.
• Key Benefits: Offers Contact Management, Feedback Collection, Multilingual Support, and Sales Assistance.
• Distinction: Unlike standard chatbots, it supports multiple Large Language Models (LLMs) including Ollama (local), OpenAI, Anthropic, and Google, allowing flexibility in performance and cost based on your needs.

Features

• Knowledge Management: Store business info in .txt files in the knowledge/ directory.
• LLM Providers: Supports Ollama, OpenAI, Anthropic, and Google Generative AI, configurable via Admin UI or environment variables.
• Semantic Search: Utilizes sentence transformers and ChromaDB for relevant query responses.
• Admin UI: Web interface (/admin) to manage knowledge, configure LLM/model, customize widget appearance, and handle backups.
• Docker Support: Deploy with Docker and Docker Compose.
• WordPress Integration: Experimental plugin structure in wordpress/chatbot/.
• Monitoring: Basic Prometheus setup included.

Installation

Docker (Recommended)

1. Clone Repository:

   git clone https://github.com/KazKozDev/ChatBot.git
   cd ChatBot

2. Configure Environment (optional): Create a .env file for API keys (see settings_manager.py for variables like CHATBOT_PROVIDER, CHATBOT_MODEL_NAME, CHATBOT_API_KEY). Defaults to Ollama if not set.
3. Build and Run:

   docker-compose up --build -d

   Server typically available at http://localhost:8000.

Manual Installation

1. Prerequisites: Python 3.9+, Ollama (if using locally), terminal access.
2. Clone Repository:

   git clone https://github.com/KazKozDev/ChatBot.git
   cd ChatBot

3. Virtual Environment:

   python3 -m venv venv
   source venv/bin/activate  # Windows: `venv\Scripts\activate`

4. Install Dependencies:

   pip install -r requirements.txt

5. Configure Environment (optional): See Docker step 2.
6. Prepare Knowledge Base: Add .txt files to knowledge/ directory for topics (e.g., about.txt, services.txt). An example.txt is created if empty on first run.
7. Run Server:

   uvicorn server:app --host 0.0.0.0 --port 8000 --reload

   Note: --reload is for development; remove for production. Ollama models are pulled if needed.

Usage

Chat Widget Integration

Add this script before </body> tag in HTML:

<script>
(function() {
    const serverUrl = 'http://localhost:8000'; // Replace with your server URL
    const script = document.createElement('script');
    script.src = serverUrl + '/static/chat-widget.js';
    script.async = true;
    document.head.appendChild(script);

    script.onload = function() {
        ChatWidget.init({
            botName: 'Assistant',
            apiUrl: serverUrl + '/chat'
        });
    };
})();
</script>

Note: Update http://localhost:8000 to your server's URL.

Admin Dashboard

Access at /admin (e.g., http://localhost:8000/admin) to:

• Configure LLM provider and model, input API keys.
• Manage knowledge files (view, edit, delete).
• Customize chat widget appearance.
• Handle backups.

Requirements

• Python 3.9+
• Docker & Docker Compose (for Docker setup)
• Ollama (for local LLM)
• Dependencies in requirements.txt
• ~2GB+ RAM (for local LLMs)

Project Structure

ChatBot/
├── knowledge/          # Knowledge base .txt files
├── static/             # Static assets (JS, CSS)
├── templates/          # HTML templates (Widget, Admin UI)
│   └── admin/          # Admin UI templates & files
├── wordpress/          # Experimental WordPress plugin
├── .env                # Environment variables (optional)
├── backup_manager.py   # Backup/restore logic
├── chatbot.py          # Chatbot logic, LLM integration
├── docker-compose.yml  # Docker configuration
├── Dockerfile          # Docker build instructions
├── knowledge_loader.py # Knowledge file processing
├── requirements.txt    # Python dependencies
├── server.py           # FastAPI server, API, Admin UI
└── settings_manager.py # Settings management

Contributing

Refer to PROJECT_PLAN.md for improvement areas. Follow existing code patterns and update documentation.

---

If you like this project, please give it a star ⭐

For questions, feedback, or support, reach out to:

Artem KK | MIT LICENSE