https://github.com/rjwignar/magic-transform
Unify diverse images into a stylish masterpiece with AI-powered style transfer!
https://github.com/rjwignar/magic-transform
canva canva-sdk
Last synced: 3 months ago
JSON representation
Unify diverse images into a stylish masterpiece with AI-powered style transfer!
- Host: GitHub
- URL: https://github.com/rjwignar/magic-transform
- Owner: rjwignar
- Created: 2024-07-04T00:50:31.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-08-26T19:46:32.000Z (over 1 year ago)
- Last Synced: 2024-12-26T10:29:46.556Z (about 1 year ago)
- Topics: canva, canva-sdk
- Language: TypeScript
- Homepage: https://www.youtube.com/watch?v=g-NNNwQHieI
- Size: 44.4 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 5
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Magic Transform - AI-Enabled Canva App for Image Generation
## Inspiration
Image generators like DALL-E are impressive, but many people struggle with
engineering effective image prompts. We wanted to create an app that allows
users to easily assemble their own images, even in disparate styles. With our
app, users can add elements to a canvas, and AI will generate a prompt from the
image to recreate it in a consistent style. This approach shifts from
text-to-image generation to a more intuitive image-to-image generation, making
image generation with AI more accessible.
## What It Does
Magic Transform allows users to create images by assembling elements on
a canvas. The app then uses AI to generate a prompt from the assembled image
and recreates it in a consistent style using DALL-E 3.
### Key Features
- **Image-to-Image Generation**: Users can add elements to a canvas, and AI
will generate a prompt to recreate the image in a consistent style.
- **Seamless Integration**: Built using the Canva SDK, OpenAI API, and
Microsoft Azure, the app provides a smooth user experience.
- **Export and Transform**: The app exports the current canvas and passes it
to OpenAI's GPT Omni model to retrieve a description, which is then used to
generate a new image with DALL-E 3.
### Technologies Used
- **Backend**: Node.js, OpenAI API, Docker, Microsoft Azure Container Apps
- **Frontend**: React, TypeScript, Canva SDK, Cloudinary
- **Deployment**: Docker containerization, Azure Container Apps, Cloudinary
for asset hosting
## What's Next
- **Codebase Improvements**: We plan to enhance the readability and
optimization of the codebase.
- **UI/UX Enhancements**: We aim to add new UI elements to improve the user
interface and user experience.
- **New Features**: We are exploring the possibility of adding features that
allow users to easily share their creations and implementing a payment model
to accept payments.