Exposure

Coach Edit Suggestions
Coach Metadata Description (lens)
Library
Manual Editing (Adjust Tab)
Automatic Editing (Generate Tab)
Output of the previous request
Basic Camera Settings

Inspiration

With cameras in everyone's phone, millions of photos are being taken every single day. We noticed that while camera technology has gotten better and more accessible, the post/editing side has been lacking in its accessibility. Most existing apps, such as Photoshop and Lightroom, are not adapted for the need of non-technical people / beginners. We wanted to create an easy to use photo editor natively designed to be integrated with an AI Copilot, an editing coach of sorts, in our case Gemini.

What it does

Our mobile editor supports basic features such as an edit history, a library, filters (various settings), layering, which can easily be controlled by both the user and the AI. Additionally, there are generative features that can create assets that seemingly transform into their own layers, which means that you can change the image using Generative AI while isolating the specific parts that changed, ensuring the fidelity of the original image. Presets also get created according to user needs, which means that if your pictures commonly come with certain flaws, you can apply fixes in one go every other time. Additionally, the user's pictures can be synced to an account (set up with Auth0), which means that you can do things from multiple devices.

How we built it

The frontend is built using Expo, a framework on top of React Native which allows us to easily build cross platform applications. Our backend is built using FastAPI, and we also depent on Auth0, MongoDB, the Gemini API, and ElevenLabs + LiveKit. We support Android and iOS out of the box, and the development environment also works on Windows, MacOS, and Linux. Most testing occured using the Android Emulator + Expo Go (an application which allows us to test Expo applications on our phones).

Challenges we ran into

Getting the development environment to run on all 3 major operating systems was difficult but necessary since our team used all 3 of Windows, MacOS, and Linux. We also could not sacrifice on either iOS or Android support, we occasionally faced operating system specific bugs related to primitives closer to the hardware, specifically the camera, which we had to address. Performance and reliability was also tough, initially the backend system was brittle and so the responses we got were questionable, but with bug fixes we were able to produce one shots (through the coach, chat, and generative tabs) easily.

Accomplishments that we're proud of

We're proud of the fact that we implemented all of the core features we initially set out for the application to have quite early on, allowing us to spend more time on ensuring that the app is as reliable as it can be.

What we learned

Mobile app development can be quite difficult, while Expo abstracts native concepts from developers as much as possible, there are still many platform-specific bugs, and so we needed to test every feature on both Android and iOS to ensure that things were stable.