This repository contains scripts for generating AI-powered videos using Luma Labs' API. The project focuses on creating smooth transitions between multiple image frames to produce captivating video sequences.
- Upload images to Bunny.net CDN for easy access (free)
- Generate videos using Luma AI with custom prompts and keyframes
- Parallel processing for efficient video generation
- Concatenate multiple generated videos into a final output
- Python 3.7+
- Luma AI API key
- Bunny.net Edge Storage account and API key
-
Clone this repository:
git clone https://github.com/Sigil-Wen/luma.git cd luma -
Install required packages:
pip install -r requirements.txt -
Set up your environment variables:
LUMA_API_KEY: Your Luma AI API keyBUNNY_STORAGE_API_KEY: Your Bunny.net Edge Storage API keyBUNNY_STORAGE_ZONE_NAME: Your Bunny.net Storage Zone name
-
Place your input images in the
./photosdirectory. -
Run the main script:
python stringvideo.py -
The script will:
- Upload your photos to Bunny.net CDN
- Generate transition videos between each pair of consecutive images
- Download the generated videos
- Concatenate all videos into a final output video
-
The final video will be saved as
final_output.mp4in the project directory.
You can modify the prompt variable in stringvideo.py to change the AI-generated content between keyframes. Experiment with different prompts to achieve various visual effects and transitions.
This project is licensed under the MIT License - see the LICENSE file for details.
generation = client.generations.create( prompt="Low-angle shot of a majestic tiger prowling through a snowy landscape, leaving paw prints on the white blanket", keyframes={ "frame0": { "type": "image", "url": "https://storage.cdn-luma.com/dream_machine/7e4fe07f-1dfd-4921-bc97-4bcf5adea39a/video_0_thumb.jpg" }, "frame1": { "type": "image", "url": "https://storage.cdn-luma.com/dream_machine/12d17326-a7b6-4538-b9b7-4a2e146d4488/video_0_thumb.jpg" } } )
