GIF
Zapped Journey
Choosing
After Chossing
Outcome
Sample Pictures

Zapped

Inspiration

Inspirated by the recent hype of text-to-image generation and the recent success of GPT-3 and DALL·E 2, we decided to create zapped, an app that helps you pick your next movie to watch using the power of AI text-to-image generation.

What it does

zappend is a web app that helps you pick your next movie to watch using the power of AI text-to-image generation. You pick between three sets of three images and the app will recommend you a movie based on your choices. The app uses the DALL·E 2 model to generate the images.

Once you have picked your movie, you can click on the movie poster to be redirected to the movie's IMDB page. You can also use up to three vetos to to get a different movie recommendation. But once you have used all your vetos, you are forced to pick the movie that the app recommends, so choose wisely!

How we built it

We built zapped using the DALL·E 2 model and the NextJS framework. We used the NextJS framework to create the frontend and backend of the app and the DALL·E 2 model to generate the images.

Challenges we ran into

Coming up with an idea that was both interesting and feasible was a challenge. We also had some issues with sleep, but we were able to fix them.

Accomplishments that we're proud of

We are proud of the app's design and the fact that we were able to get nice DALL·E 2 generated images to work with. We are also proud of the fact that we were able to deploy a fully functional app on Vercel: https://hackatum22.ospatos.com/.

What we learned

We learned how to use NextJS and Vercel to deploy a web app. We also learned how to use the DALL·E 2 model to generate images.

What's next for zapped

We would like to use a more sophisticated approach for automatically generating the images. For instance, if we would use Stable Diffusion we could encode the movie's description into a vector and then boost or weeken certain concepts (romance, violence, happy, etc.) in the vector to generate images that are more or less similar to the movie's description while adding or removing certain concepts.

-- Caro, Felix, Pia and Ricardo