HawkerSense
AI calorie tracker that actually understands Asian food
Start new thread
Qi Han Wong

HawkerSense - AI calorie tracker that actually understands Asian food

by•
Standard fitness apps are blind to Asian food. They look at Nasi Lemak, see "rice and chicken," and miss hundreds of calories in hidden coconut milk and palm oil. HawkerSense is a visual AI that decomposes mixed-plate meals component by component in 3.9 seconds. It catches the hidden saturated fats that Western trackers miss, all while judging your diet with the personality of a brutally honest Singaporean Hawker Uncle.

Add a comment

Replies

Best
Qi Han Wong
Maker
📌
Hi Product Hunt! 👋 I'm Qi-Han, the builder behind HawkerSense. I built this because tracking macros in Asia using standard fitness apps is broken. Log a plate of Nasi Lemak into a Western app, and it just sees "rice and chicken" for ~400 kcal. It completely misses the coconut milk, the hidden palm oil, and the sugar-heavy sambal. I spent my nights and weekends building an AI to fix this data gap. HawkerSense doesn't just guess the dish. It visually decomposes mixed-plate meals component by component to catch hidden fats and calories. To make the UX engaging, the AI has the personality of a brutally honest Singaporean "Hawker Uncle" who will judge your dietary choices. Traction & Tech so far: Speed: I optimized the vision model to recognize region-specific ingredients (like deep-fried tofu puffs vs. regular tofu). Average scan time is 3.9 seconds. Data: I've processed over 7,100 real-world meals on an early webapp version. The Insight: 50% of local meals scanned grade out as a C or D due to hidden saturated fats. I pushed the mobile apps live and hit #74 in Health & Fitness on the App Store today. The app offers 3 free scans per day. I'd love for the PH community to try breaking the visual AI with complex or weird food combos. I'll be here all day to answer questions about the stack, the computer vision model, or the data!