"Stop drowning in news. Start understanding it."
An intelligent full-stack web application that summarizes news articles using state-of-the-art NLP, analyzes their sentiment, reads them aloud, and tracks breaking stories over time — all behind a secure, personalized login.
🚀 Live Demo · 🐛 Report Bug · ✨ Request Feature
| 🔐 Login Page | 📝 Summarizer | 📈 News Pulse |
|---|---|---|
 |
 |
 |
💡 Replace the placeholder images above with actual screenshots from your app.
The average person encounters hundreds of news articles daily. Reading them all is impossible. Smart News Summarizer solves this by giving you the essence of any article in seconds — with AI that actually understands what it's reading.
Built during a real-world internship at Sopra Steria, Noida, this is not a toy project. It's a production-ready, full-stack application with secure authentication, real NLP models, live news APIs, and text-to-speech — deployed on Streamlit Cloud.
- Paste a URL or raw article text and get an intelligent summary instantly
- Powered by
distilbart-cnn-12-6from Hugging Face Transformers — a state-of-the-art extractive/abstractive NLP model - Choose from 3 summary lengths: One Line · Short Paragraph · Detailed
- Polarity score (Positive / Neutral / Negative) via TextBlob
- Subjectivity score (Objective / Mixed / Subjective)
- Word count and estimated reading time for every article
- Translate summaries to Hindi using Google Translate API (Deep Translator)
- More languages can be easily added
- Hit "Listen to Summary" and hear the summary read aloud via gTTS
- Works in both English and Hindi
- Enter any topic (e.g. "Climate Change", "ISRO", "Budget 2025")
- Get a 7-day chronological timeline of top headlines from NewsAPI
- Expand any headline for its description, timestamp, and source link
- Email + Password signup with mandatory email verification
- Username-based login backed by Firestore database
- Forgot Password with reset email flow
- Animated loading overlay during all auth actions
- Session-based logout with full state cleanup
smart-news-summarizer/
│
├── login.py # 🔐 Entry point — Firebase Auth (Login / Signup / Reset)
│
├── pages/
│ ├── app.py # 🧠 Main app — Summarizer + News Pulse tabs
│ └── verify.py # ✅ Email verification handler (Firebase oobCode)
│
├── .streamlit/
│ └── config.toml # 🎨 Streamlit theme config
│
├── requirements.txt # 📦 All dependencies
└── README.md
Data Flow:
User → login.py → Firebase Auth → Email Verification
→ pages/app.py → [URL Input] → Trafilatura/Newspaper3k → Text Extraction
→ HuggingFace distilBART → Summary
→ TextBlob → Sentiment Analysis
→ gTTS → Audio Output
→ Deep Translator → Hindi Translation
→ NewsAPI → News Pulse Timeline
| Layer | Technology |
|---|---|
| Frontend & Backend | Streamlit 1.47.1 |
| NLP / Summarization | HuggingFace Transformers (distilbart-cnn-12-6), PyTorch |
| Sentiment Analysis | TextBlob, NLTK |
| Text Extraction | Trafilatura, Newspaper3k, BeautifulSoup4 |
| Translation | Deep Translator (Google Translate) |
| Text-to-Speech | gTTS (Google Text-to-Speech) |
| Authentication | Firebase Auth (Pyrebase4) + Firebase Admin SDK |
| Database | Firestore (Firebase) |
| News Data | NewsAPI (/v2/everything) |
| Environment | Python-dotenv, Streamlit Secrets |
git clone https://github.com/PRERITARYA/smart-news-summarizer.git
cd smart-news-summarizerpip install -r requirements.txtCreate a .streamlit/secrets.toml file with the following:
NEWS_API_KEY = "your_newsapi_key_here"
[FIREBASE_CONFIG]
apiKey = "your_firebase_api_key"
authDomain = "your-project.firebaseapp.com"
databaseURL = "https://your-project.firebaseio.com"
projectId = "your-project-id"
storageBucket = "your-project.appspot.com"
messagingSenderId = "your_sender_id"
appId = "your_app_id"
[FIREBASE]
type = "service_account"
project_id = "your-project-id"
private_key_id = "your_private_key_id"
private_key = "-----BEGIN PRIVATE KEY-----\n...\n-----END PRIVATE KEY-----\n"
client_email = "firebase-adminsdk-...@your-project.iam.gserviceaccount.com"
client_id = "your_client_id"
auth_uri = "https://accounts.google.com/o/oauth2/auth"
token_uri = "https://oauth2.googleapis.com/token"streamlit run login.pyOpen http://localhost:8501 in your browser.
| Service | How to get it |
|---|---|
| NewsAPI | Sign up free at newsapi.org |
| Firebase | Create a project at console.firebase.google.com → Enable Email/Password Auth → Create Firestore DB → Generate Admin SDK key |
streamlit==1.47.1 # UI framework
transformers==4.54.1 # HuggingFace NLP models
torch==2.7.1 # PyTorch backend
textblob==0.17.1 # Sentiment analysis
trafilatura==2.0.0 # Article text extraction
newspaper3k==0.2.8 # Article scraping
deep-translator==1.11.4 # Google Translate API
gTTS==2.5.4 # Text-to-speech
firebase-admin==6.6.0 # Firebase Admin SDK
pyrebase4 @ git+... # Firebase client SDK
requests==2.32.4 # HTTP requests- Push your code to GitHub
- Go to share.streamlit.io
- Connect your repo — set main file to
login.py - Add all secrets in the Secrets section of the app settings
- Deploy 🎉
| 👤 Who | 💡 How they use it |
|---|---|
| Journalists & Researchers | Rapidly review dozens of articles to extract key facts |
| Business Professionals | Track competitor news and industry trends via News Pulse |
| Students | Summarize long articles for assignments and research |
| General Public | Stay informed in minutes, not hours |
This project was built during an internship at Sopra Steria India, Noida (July – September 2025).
Special thanks to Mr. Chandramouli Vashisth (Sr. Technical Lead, Sopra Steria Group) for his continuous guidance, invaluable feedback, and technical mentorship throughout the development of this project.
Prerit Arya B.Tech Computer Science & Engineering
⭐ If you found this project useful, please consider giving it a star!
Built with ❤️ during internship at Sopra Steria · 2025