NLP-powered document summarizer — paste in a PDF or text file and get a clean, extractive summary in seconds.

Briefly.CU is an intelligent document summarization web app built for students and professionals. Upload any PDF or plain text file and the app extracts the most important sentences using a TextRank-based NLP pipeline — no large language models, no API costs, fully offline.

The name is a nod to CHRIST (Deemed to be University), where this project was developed as part of MCA coursework in Natural Language Processing.

• 📂 PDF & Text Upload — supports .pdf and .txt file formats
• 🧮 Extractive Summarization — uses TextRank (graph-based ranking) to surface the most relevant sentences
• ⚙️ Adjustable Summary Length — control how many sentences appear in the output
• 🖥️ Clean Streamlit UI — simple, no-fuss interface that works in the browser
• 🔒 Fully Offline — no external API calls, all processing runs locally

User uploads PDF / TXT
        │
        ▼
  PDFPlumber extracts raw text
        │
        ▼
  NLTK tokenizes into sentences
  + removes stopwords
        │
        ▼
  scikit-learn builds TF-IDF
  similarity matrix between sentences
        │
        ▼
  NetworkX runs PageRank on
  the sentence similarity graph
        │
        ▼
  Top-N sentences extracted
  → Summary returned to UI

The core algorithm is TextRank — a graph-based ranking model inspired by Google's PageRank. Each sentence is a node; edges are weighted by cosine similarity between TF-IDF vectors. The highest-ranked nodes form the summary.

• Python 3.9+
• pip

# 1. Clone the repository
git clone https://github.com/Corerishi/Briefly.CU.git
cd Briefly.CU

# 2. Install dependencies
pip install -r requirements.txt

# 3. Download NLTK data (first run only)
python -c "import nltk; nltk.download('punkt'); nltk.download('stopwords')"

# 4. Run the app
streamlit run app.py

The app will open at http://localhost:8501

streamlit
flask
nltk
scikit-learn
networkx
pdfplumber

Developed as part of the MCA programme at CHRIST (Deemed to be University), Bengaluru, for a Natural Language Processing module. The project demonstrates practical application of graph-based NLP techniques without reliance on pre-trained transformer models.

Rishi Raj MCA — CHRIST (Deemed to be University) LinkedIn · GitHub

This project is licensed under the MIT License.