Inspiration

I have always wanted to build an image-to-video AI app and never released an app to the app store before, so I decided to make a paid app that allows users to generate cool AI videos based on their images.

What it does

Transform any image or prompt into a cinematic video, either by using our custom templates or by creating your own. You can also transform photos into AI photos by using our provided templates.

How we built it

React Native on the client and Node.js/Express on the backend.

Challenges we ran into

Building the engine behind the generations was quite challenging; we tried many AI APIS, but none did well with facial recognition, and especially building templates was quite tricky. Scaling is also quite challenging since we do some post-video processing on our servers, so we need to scale if we get a lot of concurrent generations (especially for 1080p generations)

Accomplishments that we're proud of

We worked some magic and were able to build really awesome templates that also preserve your face!

What we learned

A lot about AI prompt generation for videos and photos, hacks and tricks around perfecting outputs, building predefined templates, creating tasks, etc..

What's next for PixMorph AI

Build a credits system ledger + implement distributed horizontal scaling solutions on our servers + add more templates + release on Android too

Built With

Share this project:

Updates