📝 Summify

A lightweight text summarization web app built with Streamlit and Hugging Face Transformers. Paste a paragraph or article and get a concise summary generated by the pretrained facebook/bart-large-cnn model.

Features

• Abstractive summarization using a BART transformer
• Simple, clean Streamlit interface
• Model loaded once and cached with st.cache_resource, so repeat summaries are fast
• Friendly warning when the input is empty
• Deterministic output (do_sample=False), with summaries of 30-130 tokens

Demo

Tech stack

Component	Purpose
Streamlit	Web interface
Hugging Face Transformers	Summarization pipeline
PyTorch	Model backend
facebook/bart-large-cnn	Pretrained summarization model

Getting started

Requirements: Python 3.9+

git clone https://github.com/<your-username>/summify-text-summarizer.git
cd summify-text-summarizer

python -m venv venv
venv\Scripts\activate        # Windows
# source venv/bin/activate   # macOS / Linux

pip install -r requirements.txt
streamlit run app.py

The app opens at http://localhost:8501.

Note: The first run downloads the model (about 1.6 GB), so startup will be slow once. After that it loads from the local cache.

How it works

Pasted text
     ↓
Hugging Face summarization pipeline (BART)
     ↓
Summary (30-130 tokens)
     ↓
Displayed in Streamlit

Core of the app:

summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
summary = summarizer(text, max_length=130, min_length=30, do_sample=False)[0]["summary_text"]

Limitations

• BART accepts roughly 1,024 tokens of input. Very long text may be truncated or cause an error.
• The model is trained mainly on news articles (CNN/DailyMail), so it works best on news-style text.
• Runs on CPU by default, so long inputs can take several seconds.
• Summaries are abstractive and can occasionally contain inaccuracies, so check important facts against the source.

Roadmap

• Split long documents into chunks and summarize each
• Adjustable summary length slider
• Upload .txt / .pdf files
• Option to choose between different summarization models
• Deploy on Streamlit Community Cloud or Hugging Face Spaces

Project structure

summify-text-summarizer/
├── app.py
├── requirements.txt
├── README.md
└── docs/
    └── screenshot.png

License

MIT (add a LICENSE file when you create the repo).