Sentiment Analyzer Web Extension

A powerful Chrome extension that uses AI to instantly analyze the emotional tone of any text on the web. Powered by a fine-tuned DistilBERT model, it detects positive, negative, or neutral/mixed sentiments with confidence scoring.

Features

• Instant Analysis: Right-click any selected text and analyze its sentiment in seconds
• AI-Powered: Uses a fine-tuned DistilBERT model optimized for sentiment classification
• Confidence Scoring: Displays confidence levels with visual progress bars
• Clean UI: Modern, animated interface with color-coded sentiment indicators
• Privacy-Focused: Text is only processed when you explicitly request it
• Fast Response: Optimized API hosted on Hugging Face Spaces

How It Works

1. Select any text on a webpage
2. Right-click and choose "Analyze Sentiment"
3. View instant results with sentiment classification and confidence score

Tech Stack

Extension (Frontend)

• Chrome Manifest V3
• Vanilla JavaScript
• CSS3 with smooth animations
• Chrome Extension APIs

Backend API

• Framework: Flask
• Model: Fine-tuned DistilBERT (3-class sentiment classifier)
• ML Library: Transformers (Hugging Face)
• Deployment: Hugging Face Spaces
• Server: Gunicorn

Installation

Installing the Extension

1. Clone the repository:

   git clone https://github.com/Destroyer795/SA-Web-Extension.git
   cd SA-Web-Extension

2. Load in Chrome:

   • Open Chrome and navigate to chrome://extensions/
   • Enable Developer mode (toggle in top-right)
   • Click Load unpacked
   • Select the extension folder from this project

3. Start using:

   • Select any text on a webpage
   • Right-click and choose "Analyze Sentiment"

Running the API Locally

1. Install dependencies:

   pip install -r requirements.txt

2. Run the Flask server:

   python app.py

3. Update API endpoint (if running locally):

   • Open extension/popup.js
   • Change API_URL to http://localhost:7860/predict

Configuration

Model Details

• Model Type: DistilBERT for Sequence Classification
• Classes: Positive, Negative, Neutral
• Confidence Threshold: 70% (scores below are marked as "Neutral / Mixed")
• Location: ./sentiment_analyzer_pro/

API Endpoints

POST /predict

Analyzes sentiment of provided text.

Request:

{
  "text": "This is an amazing product!"
}

Response:

{
  "sentiment": "Positive",
  "score": 0.9542,
  "confidence_flag": "High"
}

GET /

Health check endpoint.

Project Structure

.
├── extension/              # Chrome extension files
│   ├── manifest.json      # Extension configuration
│   ├── popup.html         # Popup UI
│   ├── popup.js           # Popup logic
│   ├── popup.css          # Popup styles
│   └── background.js      # Background service worker
├── sentiment_analyzer_pro/ # Fine-tuned DistilBERT model
│   ├── model.safetensors
│   ├── config.json
│   ├── tokenizer.json
│   └── ...
├── app.py                 # Flask API server
├── requirements.txt       # Python dependencies
├── Dockerfile            # Container configuration
└── README.md

Sentiment Classification Logic

The model returns a label and confidence score:

• High Confidence (≥70%): Returns the predicted sentiment (Positive/Negative)
• Low Confidence (<70%): Returns "Neutral / Mixed" to handle ambiguous or sarcastic text

This approach reduces false positives for edge cases like sarcasm or mixed emotions.

Deployment

Hugging Face Spaces

The API is currently deployed at:

https://destroyer795-sentiment-analyzer-extension.hf.space/

To deploy your own:

1. Create a Hugging Face Space
2. Upload the project files
3. Update the API_URL in extension/popup.js

Docker

Build and run using Docker:

docker build -t sentiment-analyzer .
docker run -p 7860:7860 sentiment-analyzer

Contributing

Contributions are welcome! Feel free to:

• Report bugs
• Suggest new features
• Submit pull requests

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Hugging Face for the Transformers library and hosting platform
• DistilBERT model architecture by Hugging Face
• Chrome Extension APIs documentation

Contact

• Developer: Destroyer795
• Hugging Face Space: Sentiment-Analyzer-Extension