Skip to main content

My contribution

I built the frontend and backend, including profile setup, progress tracking, account flows, and the Hale guidance interface.

How I worked on it

  • Separated deterministic calculations from optional online AI so the main application does not depend on an external model.
  • Built guest and account paths with different storage rules, then added charts, nutrition workflows, and verification around those paths.
HaleView: HaleView dashboard with synthetic guest data.
HaleView dashboard with synthetic guest data.

What it is

HaleView brings profile setup, activity records, progress charts, nutrition planning, and the Hale assistant into one application. Core calculations and local guidance work without online AI. Guest mode keeps the profile in the browser; account mode uses a backend and SQLite.

Features

  • Guest and account access
  • Profile setup, records, and progress charts
  • Nutrition and shopping workflows
  • Local guidance with optional online AI
  • Account verification, session refresh, and optional two-step sign-in

Stack

TypeScriptReactExpressSQLiteDocker

How to run it

git clone https://gitea.kood.tech/alekseyaleksandrovich/numbers-dont-lie.git
cd numbers-dont-lie
docker compose up --build
# Open http://127.0.0.1:27450