HackStack

Project Info

CalTracker

Devpost

Inspiration

We’re fitness-focused people who care about what we eat and how we perform. Tracking nutrition was a daily habit, but manual entry was slow and inaccurate. We wanted a smarter way—one that could estimate nutritional data directly from food images while also helping people stay accountable through social features.

What it does

CalTracker lets users authenticate, upload images of their meals, and get detailed food information automatically using Google Gemini’s image analysis. The app stores nutrition stats tied to each user and allows friends to form social groups to track progress together. It’s a mix of AI-powered food recognition, fitness tracking, and community motivation.

How we built it

Backend: FastAPI (Python) for APIs and data management. Frontend: React for the web interface. Mobile: Swift for iOS. Dashboard: Svelte + Tauri for a lightweight, cross-platform analytics dashboard. The AI pipeline integrates directly with Google Gemini for image-based nutrition estimation. User authentication and data storage tie all the components together into a cohesive system.

Challenges we ran into

CORS issues caused major headaches—cross-site requests between our various frontends and backend needed constant debugging. Coordinating multiple frameworks (React, Svelte, Swift, FastAPI) was tough, especially with limited frontend experience. Managing so many moving parts introduced technical debt and integration pain points.

Accomplishments we're proud of

We aimed high and delivered something real. We tackled multiple unfamiliar technologies, built a complex multi-platform system, and made something that could actually improve people’s health. The ambition itself was a success: connecting AI, nutrition, and social accountability into one product.

What we learned

How CORS works and how fragile it can be. The cost of technical debt when using AI-assisted coding. Time management and the importance of realistic scope. The value of teamwork and learning outside your comfort zone. We learned that growth happens at the edge of frustration.

What's next

We’re done with this iteration—it served its purpose as a learning project. The next step is to build something better: cleaner architecture, tighter design, stronger purpose. CalTracker was a stepping stone toward future systems that are better built, better designed, and help more people. [ \text{Growth} = \frac{\text{Experience}}{\text{Comfort Zone}} ]

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • CSSIn code
  • FastAPIIn code
  • HTMLIn code
  • JavaScriptIn code
  • PythonIn code
  • ReactIn code
  • SupabaseIn code
  • SvelteIn code
  • SwiftIn code
  • Tailwind CSSIn code
  • TypeScriptIn code
  • Google GeminiClaimed
  • Node.jsClaimed
  • PostgreSQLClaimed

11 of 14 appear in the indexed code. 3 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

101 KB

Source files

44

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars