SproutBowl: AI Family Meal Tracker
Privacy-first, AI-powered meal tracker for the whole family.
Start new thread
Ikuya Terasaka

SproutBowl: AI Family Meal Tracker - Privacy-first, AI-powered meal tracker for the whole family.

by•
Track your family’s nutrition in seconds. Snap a photo, and our on-device AI instantly categorizes meals into 9 essential groups. Perfect for toddler meals and allergy management. 100% private, no photos leave your device. Complete free & no limits.

Add a comment

Replies

Best
Ikuya Terasaka
Maker
📌
Hi Product Hunt community! 👋 I’m the maker of SproutBowl. As a parent and a developer, I always struggled with existing meal trackers. They are mostly built for individual calorie counting and dieting, making it incredibly tedious to manage a whole family's daily nutrition. Who has time to manually log every single ingredient while rushing through dinner with toddlers? That’s why I built SproutBowl. Here is what makes it different: 1. Family-Centric (Today's Plate): Visualize the whole family’s nutritional balance across 9 vital food categories in one clean dashboard. 2. 1-Tap On-Device AI: Just snap a photo. Our offline AI categorizes the meal instantly. 3. Strict Privacy: Because it runs completely on-device, your family photos never touch a server or leave your phone. 4. Parent Tools: Includes allergy tracking and a fun kids mode to make healthy eating engaging. The app is completely free with no usage limits. I’d love to hear your thoughts, feedback, or any feature requests you have for your own family's meal routines! Thank you so much for checking it out! 🍲✨
Nicole Hynek

@kichiemon Love the privacy-first approach, especially for families sharing photos of meals and children. Making nutrition tracking as simple as taking a photo removes a lot of the friction that causes people to quit. How accurate is the categorization when meals contain multiple ingredients or mixed dishes?

Ikuya Terasaka
Maker

@nicole_hynek Thank you for your feedback. I'm glad our approach to privacy and reducing friction resonates with you.

Regarding the accuracy for meals with multiple ingredients or mixed dishes: we are currently using a custom-trained model that runs entirely on-device. This is also how we ensure the privacy aspect.

To be transparent, the accuracy for highly complex mixed dishes is still a work in progress. While the model handles primary ingredients reasonably well, it can sometimes struggle to identify every single item in a mixed dish. We are continuously training and refining the model to improve its categorization accuracy.