🗡️ Summaraii

Cut the Clutter, Keep the Core.

Summaraii is a Streamlit app that helps creators quickly extract topic-relevant insights from YouTube videos, websites, and PDFs. It filters each source by your topic first, then summarizes only what matters using an LLM backed by Groq.

Production demo: https://summaraii.streamlit.app/

Features

• Multi‑source input: YouTube URLs, website URLs, and PDFs.
• Topic filtering first: Content is filtered by your topic before summarization.
• LLM‑powered summaries: Uses Groq Gemma2‑9b‑It via LangChain.
• Clean UI: Streamlit app with a two‑column layout and status messaging.

Quickstart

Prerequisites:

• Python 3.9+
• A Groq API key

Clone and install:

git clone https://github.com/<your-org-or-user>/Summaraii.git
cd Summaraii
python -m venv .venv && .venv\Scripts\activate  # on Windows PowerShell: .venv\Scripts\Activate.ps1
pip install -r requirements.txt

Configure environment:

• In the app sidebar you can paste your key each run, or create a .env file with:

GROQ_API_KEY=your_groq_api_key

Run the app:

streamlit run cs.py

Open http://localhost:8501 in your browser.

Usage

• Enter a topic in the sidebar.
• Optionally add YouTube and website URLs (one per line) and/or upload PDFs.
• Click “Summarize Content”. The app filters by topic and displays a combined summary.

Configuration

• Model: Gemma2-9b-It via langchain-groq.
• Chunking: RecursiveCharacterTextSplitter with 5000/500.
• Filtering: simple regex match of your topic (case‑insensitive) before summarization.

Development

Recommended commands:

# Lint (optional: add flake8/ruff to your env first)
python -m pip install ruff flake8
ruff check . || true
flake8 || true

# Run Streamlit
streamlit run cs.py

Contributing

Contributions are welcome! Please read CONTRIBUTING.md and follow the code of conduct in CODE_OF_CONDUCT.md. For security issues, see SECURITY.md.

License

MIT License. See LICENSE.