A hybrid chatbot architecture combining a Python-based NLP engine with a lightweight JavaScript frontend for seamless web integration.
Developed by Yooungs Studios
KinglyShade V2 represents the evolution of our conversational agents. Unlike standard rule-based bots, this system utilizes a custom-trained machine learning model to classify user intents and generate context-aware responses.
The architecture is designed to be decoupled: the AI logic resides in a robust Python backend (app.py), while the interface is served via a responsive HTML/JS layer, allowing for easy integration into existing web platforms.
The system operates on a client-server model optimized for low latency response times.
| Component | Technology | Description |
|---|---|---|
| NLP Engine | Python / TensorFlow | Handles tokenization, intent classification, and confidence scoring. |
| Data Layer | JSON Structure | Manages training data (intents.json) defining user patterns and responses. |
| Frontend | Vanilla JS / CSS3 | Asynchronous UI that communicates with the backend API via Fetch/XHR. |
| Model Loader | Custom Script (cargar.py) |
Efficiently loads pre-trained .h5 or .pkl models into memory. |
KinglyShadeV2/
├── models/ # Serialized ML models and weights
├── generated-images/ # Output directory for visual assets
├── img/ # Static assets and UI resources
├── app.py # Main application entry point (Server)
├── cargar.py # Model loading utility
├── intents.json # Training dataset and responses
├── index.html # Main user interface
├── script.js # Frontend logic & API communication
└── styles.css # Responsive design rules
Instead of relying on third-party APIs (like Dialogflow), KinglyShade runs its own classification engine. This ensures data privacy and allows for unlimited customization of intents via the intents.json file.
The bot is capable of handling complex interactions by combining:
- Deterministic Responses: For precise FAQs defined in the dataset.
- Probabilistic Matches: Using the trained model to understand variations in user input.
The frontend is built without heavy frameworks, ensuring a <50kb footprint. It can be embedded into any web application by simply referencing the main script and stylesheet.
- Python 3.8+
- Pip (Python Package Manager)
-
Clone the repository
git clone [https://github.com/Yooungs-Studios/KinglyShadev2.git](https://github.com/Yooungs-Studios/KinglyShadev2.git) cd KinglyShadev2 -
Install Dependencies (Ensure you have a virtual environment active)
pip install tensorflow flask nltk
-
Run the Engine
python app.py
-
Launch Interface Open
index.htmlin your preferred browser or serve it via a local web server.
Proprietary Software. This project was developed by Yooungs Studios. The source code is provided for portfolio demonstration purposes only.
- Commercial use without license is prohibited.
- External contributions are currently closed.