Calora AI

• 

  Banner

• 

  Dashboard

• 

  Overview

Inspiration

Tracking calories should feel natural, not like filling spreadsheets. Most fitness apps require typing, scrolling, and manual logging. I wanted to build something effortless, I use chatGPT everyday to log my macros.

What it does

Calora is a calorie and macro tracker. You simply type: “I had oatmeal and a protein shake,” and Calora logs your intake, calculates nutrients, and updates your daily progress.

How we built it

I integrated Gemini API to understand user input, an NLP parser for extracting calories and macros, and backend database to store progress. The prototype runs on Python with Flask for backend logic.

Challenges we ran into

Mapping natural language to structured nutritional data required balancing between user freedom and predictable parsing.

Accomplishments that we're proud of

I built a functional prototype that allows a full calorie-tracking conversation that is powered by an LLM. It feels natural, and provides instant, feedback. The integration between multiple APIs works seamlessly in real time.

What we learned

I learned how to synchronize NLP in a single feedback loop while keeping latency low. Designing a natural conversation flow for a nutrition app also taught me how important phrasing, confirmation, and pacing is.

What's next for Calora

I plan to expand Calora AI with automatic food recognition from voice and photos, personalized dietary insights, smartwatch integration, and multilingual support. The long-term goal is to make Calora AI a fully conversational health companion that learns from user habits to provide smarter, proactive nutrition guidance.

Built With

• auth0
• elevenlabs
• flask
• gemini
• sqlite