Share

APRIL 9, 2025

Fast, accurate nutritional analysis with CalCam and Gemini 2.0 Flash

Vishal Dharmadhikari

Product Solutions Engineer

Queena Qiu

Co-founder of Polyverse

AgentOps showcase hero

The Gemini API enables developers to create complex applications that feel simple and intuitive for end users. App developer Polyverse helps users track their nutrition through the CalCam app, which uses AI to analyze images for caloric intake and power engaging features such as nutrition posters and meal ratings. By using the Gemini API, Polyverse improved response times, increased user satisfaction ratings by 20%, and enhanced internal workflows with streamlined tooling.

The CalCam app analyzes the user’s photo of a meal, calculates macronutrients, and tracks caloric intake.

Enhanced precision for better health insights

CalCam users track their nutrition by providing photos of their meals for analysis. After processing the photo, Gemini 2.0 Flash generates structured JSON outputs that integrate with CalCam’s workflow to calculate dish ingredients, weights, and macronutrient information. These outputs are further assessed against nutritional knowledge and logic to enhance the accuracy and consistency of the results.

CalCam users can provide corrections if needed, prompting the Gemini model to re-evaluate and generate a new, refined analysis.

Integrating Gemini 2.0 Flash’s abilities to pick up on subtle elements, such as sauces and seasonings, and refine analyses based on feedback, CalCam empowers users to track their meals with better precision and make informed decisions about their diet and health.

Streamlined app development with the help of Google AI Studio

Polyverse highlights the ease of customizing Gemini models as critical to accelerating the development process. The structured output visual editor in Google AI Studio allowed non-programmers on the CalCam team to contribute to structuring and editing outputs, reducing the reliance on coding expertise.

Google AI Studio’s structured output editor can be used to specify the JSON schema for model responses.
Note: Data is an example schema for visualization purposes only and not representative of real app data.

In the future, Polyverse plans to develop even more interactive and personalized features, such as AI-driven recipes and coaching, to achieve CalCam's mission of making healthy living fun and accessible. By leveraging powerful multimodal reasoning and structured output tools in the Gemini API, app functionality will continue to improve and make it even more valuable for health-conscious users.

Download CalCam to start working towards your nutrition goals and explore image reasoning with Gemini 2.0 Flash in Google AI Studio.