1 · The problem
School students face a flood of diet misinformation on social media. Food content is full of fear, fads and hidden ads. Meanwhile, food apps rank restaurants with black-box scores nobody can verify. Kids need to learn how to judge food themselves — not just obey a score.
Misinformation
Viral diet myths travel faster than facts. Nobody teaches kids how to verify nutrition claims.
Black-box ratings
Big food apps never reveal why a place scored well — you must simply trust it.
Fear-based learning
Most “diet” apps use timers, streaks and guilt. For vulnerable students, that pressure backfires.
2 · The solution
3 · The transparent Health Score
Inspired by the Michelin Guide, which earns trust by publishing its criteria. Every food gets a score out of 100 built from 5 honest criteria — and a “Why this rating” button on everything.
Example food scores
| Food | Category | Health Score | Why (1 line) |
|---|---|---|---|
| 🫓 Idli + sambar | Everyday | 82/100 | Steamed + fermented + lentil protein, low oil. |
| 🍛 Dal + roti + sabzi | Everyday | 85/100 | The balanced plant-protein plate. |
| 🥞 Masala dosa | Sometimes | 64/100 | Fermented, but potato–ghee filling adds oil. |
| 🍕 Veggie pizza | Sometimes | 58/100 | Thin crust + veg can be OK; cheese & crust sink it. |
| 🥟 Samosa | Treat | 42/100 | Fried maida shell soaked in oil. |
4 · Locate Nearby — real data, honest rating
The app fetches real eateries from OpenStreetMap within 2 km of the user's PIN code (or a safe one-time location lookup). Each venue gets a rating from 100 on 4 transparent criteria — and every rating can be opened to see its own breakdown.
5 · Safety by design (the ethics layer)
Health Quest was designed after studying three case studies, and every lesson maps to a concrete design choice.
| Case study | What we learned | How Health Quest applies it |
|---|---|---|
| Pokémon Go | Location features can cause real-world harm. | Consent-first flow, no silent tracking, PIN-code alternative. |
| Blue Whale challenge | Manipulative mechanics harm vulnerable kids. | No timers, no streaks, no “you must go”. |
| Michelin Guide | Trust comes from transparent criteria. | Visible score breakdown + “Why this rating” button. |
6 · Technology (all free)
Single HTML file · vanilla JavaScript · no framework · works offline from a plain file.
Local food database + OpenStreetMap Overpass API via a serverless proxy.
OpenRouter → NVIDIA NIM → offline fallback. Bring Your Own Key, stored in localStorage.
Netlify or Vercel free tier with a 20-line serverless function.
MIT License · K Subramanya Satyanarayana · St. John's English Medium School.
Aligned with IEEE 7000 (ethics in design) and IEEE 2651xx. See the IEEE configuration page →
7 · Privacy promise
No API keys in the HTML (BYOK only). No backend database. No user tracking, no analytics, no cookies. Location used once and never stored. Keys live in localStorage and vanish when browser data is cleared.
8 · Roadmap
| Idea | Effort | Benefit |
|---|---|---|
| Telugu translation | Low | Local relevance for Andhra Pradesh. |
| Offline PWA (service worker) | Medium | Works with no internet after first load. |
| Favourite eateries (localStorage) | Low | Personal, private, zero backend. |
| “Report outdated info” | Medium | Crowd-verified data quality. |
| School-level leaderboard (no names) | Medium | Light competition, zero pressure. |
9 · Credits & open data
- OpenStreetMap + Overpass API — free, open geodata
- Nominatim — free PIN-code geocoding
- OpenRouter + NVIDIA NIM — free AI tiers
- Netlify / Vercel — free static hosting with serverless functions
- Case-study inspiration — Pokémon Go, Blue Whale challenge, Michelin Guide