Distill — AI Content Summarizer

A Flask web app that summarizes YouTube videos and web pages into clean, numbered English summaries using Groq LLMs and LangChain.

---

Features

• YouTube support — auto-detects the URL, fetches the transcript (with language fallback and auto-translation to English)
• Website support — scrapes and cleans article/page content
• Chunk-based summarization — handles long content via a two-pass map-reduce pipeline
• Model selection — choose between Groq's available LLMs
• English-only output — always produces a structured English summary
• Dark editorial UI — polished, responsive, no frameworks required

---

Project Structure

summarizer/
├── app/
│   ├── __init__.py          # Flask app factory
│   ├── config/
│   │   └── settings.py      # Config class (reads .env)
│   ├── routes/
│   │   └── main.py          # Flask routes (GET / and POST /summarize)
│   ├── services/
│   │   ├── youtube_service.py    # YouTube transcript loading
│   │   ├── website_service.py    # Web content loading
│   │   └── summarizer_service.py # LLM chain + chunking
│   ├── utils/
│   │   └── url_helpers.py   # URL validation and parsing
│   ├── templates/
│   │   └── index.html       # Main page (Jinja2)
│   └── static/
│       ├── css/style.css    # Design system
│       └── js/app.js        # Frontend logic
├── run.py                   # Entry point
├── requirements.txt
├── .env.example
├── .gitignore
├── Data/
├── research/
├── asset/
└── README.md

---

Setup

1. Clone or copy the project

cd YT-Video-Summarizer

2. Create a virtual environment

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

3. Install dependencies

pip install -r requirements.txt

4. Configure environment

cp .env.example .env
# Edit .env — set SECRET_KEY at minimum

5. Run the app

python run.py

Open http://localhost:5000 in your browser.

---

Usage

1. Paste a YouTube URL or any public webpage URL
2. Enter your Groq API key (get one free at console.groq.com)
3. Select a model
4. Click Summarize

The app will:

• Load the content (transcript or web page)
• Split it into overlapping chunks
• Run each chunk through a first-pass bullet summarizer
• Combine all partials and produce a final structured English summary

---

Getting a Groq API Key

1. Go to console.groq.com
2. Sign up or log in
3. Generate an API key under API Keys

---

Environment Variables

Variable	Required	Description
SECRET_KEY	Yes	Flask session secret (set a long random string in production)
FLASK_DEBUG	No	Set true for dev auto-reload (default: false)

---

Extending

• Add a new model: update GROQ_MODELS in app/config/settings.py
• Change chunk size: adjust CHUNK_SIZE / CHUNK_OVERLAP in settings
• Add a new route: create a file in app/routes/ and register the blueprint in app/__init__.py
• Add output language: modify the prompts in summarizer_service.py

---

Tech Stack

Layer	Technology
Backend	Python 3.11+, Flask 3
LLM	Groq API via langchain-groq
Chains	LangChain Core + Text Splitters
YouTube	youtube-transcript-api
Web scraping	langchain-community UnstructuredURLLoader
Frontend	Vanilla HTML/CSS/JS (no framework)
Fonts	Playfair Display, DM Sans, DM Mono

Screenshots

Here are some previews of the application interface and core features.

1. Landing Page

Shows the main interface where users can enter a YouTube or website URL.

2. Input Section

Users can provide the URL, choose the model, and start the summarization process.

3. Summary Result

Displays the generated English summary in a clean and readable format.