Inspiration
The idea for Calorific came from my own struggles with nutrition tracking. I found it incredibly time-consuming to manually input each ingredient, especially when I wasn't sure what was in my meal, like when dining out. I wanted a solution that would make this process quick and effortless, and that’s where the inspiration for Calorific was born.
What it does
Calorific makes tracking your nutrition a breeze. Just open the app, snap a photo of your meal, and get instant nutritional insights. The app combines artificial intelligence with accurate nutrition information from barcodes to calculate your meal's nutritional content. No more manual entry or guesswork – simply snap a picture and let Calorific handle the rest.
How I built it
I built Calorific using Swift and SwiftUI for the frontend, and Firebase for the backend. The journey started with prototyping the AI in a playground to see if it was capable of accurately estimating nutritional information. Once I validated the concept, I created a basic MVP and then began the process of refining and polishing the app.
Challenges I ran into
One of the biggest challenges was finding the right balance between accuracy and simplicity. For example, serving sizes were often unclear, leading to inconsistent results from the AI model. While asking users to fill in serving sizes would improve accuracy, I wanted to keep the app straightforward and easy to use. To solve this, I leveraged data from barcodes, fine-tuned the AI model, and allowed for optional user input. This approach helped maintain the app's simplicity while ensuring it remained accurate.
Accomplishments that I'm proud of
I'm incredibly proud of how Calorific simplifies nutrition tracking. Users can just open the app, snap a picture, and get all the nutritional information they need – it’s that simple. My goal was to create an intuitive, native app that’s both easy to use and enjoyable to interact with, and I believe I’ve achieved that. Additionally, I'm especially proud of how I used the latest technologies like computer vision and AI to genuinely solve a problem, rather than using them just for the sake of it. This thoughtful integration of technology and design truly sets Calorific apart.
What I learned
Throughout this project, I learned that AI is inherently non-deterministic, meaning it won’t always produce consistent results. Instead of fighting this, I found ways to work around it to suit my use case. I also discovered that maintaining simplicity is crucial; it’s easy to add extra screens, but keeping the user’s perspective in mind and ensuring a smooth flow is far more challenging and important.
What's next for Calorific - Nutrition Tracker
A lot! Since the current version is an MVP, I focused on the core features and saved many ideas for future updates. You can look forward to tighter integration with iOS and improved accuracy of the model, including these features:
- HealthKit integration
- iCloud syncing
- Dictation support
- Widgets for quick entry
- Improved model estimation using depth information from the camera
Log in or sign up for Devpost to join the conversation.