LYRA - LYRics Animated

LYRA is an AI-powered music video generator that creates animated visuals synchronized with song lyrics.

Features

• Lyrics Fetching: Automatically fetches synced lyrics (LRC) from LRCLIB.
• AI Image Generation: Generates artistic, hand-drawn style illustrations for each segment using Google Gemini.
• AI Video Animation: Animates these illustrations into video clips using Odyssey.
• Video Stitching: Combines all segments into a final music video.
• Web Interface: Pure Django-based web interface - no Node.js required!
• Job Management: Cancel pending or in-progress video generations with one click.
• Real-time Updates: Auto-refreshing job status every 3 seconds.
• Dark Mode: Beautiful dark theme with smooth transitions and persistent preference.
• Smart Search: Real-time search suggestions with synced lyrics indicators.
• Visual Progress: See images and lyrics segments appear in real-time as they are generated.

Quick Setup (Conda)

1. Ensure you have Conda installed.

2. Create the environment:

   conda env create -f environment.yml

   Or run:

   bash scripts/create_conda_env.sh

3. Activate the environment:

   conda activate odyssey-hackathon

4. Important for macOS users: If you encounter Pillow/libtiff errors, run:

   brew install libtiff
   ln -sf /opt/homebrew/Cellar/libtiff/4.7.1/lib/libtiff.6.dylib /opt/homebrew/lib/libtiff.5.dylib

Running the Application

Start the Django Server

Navigate to the backend directory and run:

cd backend
./start_server.sh

Or manually with:

cd backend
export DYLD_LIBRARY_PATH=/opt/homebrew/lib:$DYLD_LIBRARY_PATH
python manage.py runserver 8000

Access the Application

Open your browser and go to: http://localhost:8000

That's it! The application now uses Django templates - no Node.js required.

Features

• Search for Songs - Enter song name and artist to find lyrics
• Generate Videos - Click "Generate Video" to create AI-powered music videos
• Monitor Progress - Watch real-time progress updates (auto-refreshes every 3 seconds)
• View Videos - Play and download completed videos directly in the browser

API Endpoints (Optional)

The REST API is still available for programmatic access:

• GET /api/jobs/ - List all video generation jobs
• POST /api/jobs/ - Create a new video generation job
• GET /api/jobs/search/?q=<query> - Search for song lyrics
• GET /admin/ - Django admin interface

Quick Setup (venv)

1. Install Python 3.12+
2. Create and activate a virtual environment:

python3.14 -m venv .venv
source .venv/bin/activate

1. Install dependencies:

pip install -e .

1. Or run:

bash scripts/create_venv.sh

Project dependencies and Python version are managed in pyproject.toml.

The Odyssey Python SDK will be installed automatically from GitHub:

• https://github.com/odysseyml/odyssey-python If you want to install it manually:

The Google Generative AI SDK (`google-genai`) will also be installed automatically.
To install it manually:
```bash
pip install -q -U google-genai

The Pillow library (`PIL`) will also be installed automatically.
To install it manually:
```bash
pip install Pillow

pip install git+https://github.com/odysseyml/odyssey-python.git

Note: Make sure to update the GOOGLE_API_KEY export GOOGLE_API_KEY=<API_KEY>

test_nano_banana.py -> Works with Google API Key.