Frostbyte Hackathon 2026 — Sustainability & Climate Tech
Scan receipts. Track your impact. Save the planet — one purchase at a time.
| Dashboard with Live Analytics | Transport Emissions Comparison |
|---|---|
 |
 |
| Food & Grocery Tracker | Learn & Discover |
|---|---|
 |
 |
| Feature | Description |
|---|---|
| 📸 Scan Receipts | Upload a grocery receipt photo — Claude Vision extracts items and calculates carbon footprint instantly |
| 🛒 Manual Tracking | Search or browse 80+ food items with real emission data from EPA/DEFRA |
| 🚗 Transport Tracker | Log commutes across 15 transport modes with side-by-side comparisons |
| 🏠 Home Energy | Track electricity & heating by source (including Quebec's near-zero hydro grid!) |
| 🤖 AI Climate Coach | Personalized, encouraging tips from Claude — data-driven, never preachy |
| 📈 Visual Dashboard | Interactive Plotly charts: category breakdown, item comparison, country benchmarks |
| 📚 Learn | Educational content with food comparisons, top 5 actions, and data sources |
People can't change what they can't measure.
Most people want to reduce their carbon footprint but have no idea where to start. ClimaByte bridges the knowledge gap:
- 👁️ See the impact of every purchase in kg CO₂e
- 🔄 Compare alternatives — "What if I swapped beef for chicken?"
- 🤖 Get coached by AI that's encouraging, not judgmental
- 📊 Track progress with beautiful, real-time visualizations
- 🌍 Benchmark yourself against global and country averages
| Layer | Technology | Why |
|---|---|---|
| Language | Python 3.12 | Massive AI ecosystem, hackathon-proven |
| Frontend | Streamlit 1.55 | Production-quality web UI from pure Python |
| AI Brain | Claude API (claude-sonnet-4-6) | Multimodal vision + structured JSON + natural coaching |
| Carbon Data | Built-in DB (80+ items) | Real emission factors from EPA, DEFRA, Climatiq OEFDB |
| Charts | Plotly Express | Interactive, beautiful, responsive visualizations |
| Image Processing | Pillow (PIL) | Receipt photo handling before Claude Vision |
| Deployment | Streamlit Community Cloud | Free, instant, auto-deploy from GitHub |
Total cost: $0 — Everything runs on free tiers.
# Clone
git clone https://github.com/limem01/climabyte.git
cd climabyte
# Setup
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
pip install -r requirements.txt
# Run
streamlit run app.pyOpens at http://localhost:8501 ✨
| Key | Required? | What it unlocks |
|---|---|---|
| Anthropic Claude | Optional | Receipt scanning (Vision) + AI coaching |
| Climatiq | Optional | Enhanced emission data |
The app works fully without any API keys using the built-in emission factor database.
ClimaByte was built for the Frostbyte Hackathon 2026. The core insight: people can't change what they can't measure.
📸 Receipt Photo → Claude Vision → Structured JSON → Emission Database Match → CO₂e Calculation
↓
🛒 Manual Entry ────────────────────────────────────────→ Carbon Engine ──→ Dashboard + Charts
↑ ↓
🚗 Transport / 🏠 Energy ──────────────────────────────────────┘ 🤖 AI Coach Tips
Key decisions:
- Built-in emission data (not API-only) so the app works even offline
- Claude claude-sonnet-4-6 for the perfect balance of speed, intelligence, and cost
- Streamlit for rapid prototyping — professional UI in hours, not weeks
- Session-based storage for hackathon simplicity (Firebase/Supabase planned for v2)
All emission factors come from peer-reviewed, authoritative sources:
- Climatiq OEFDB — Open Emission Factor Database
- EPA — US Environmental Protection Agency
- DEFRA — UK Dept. for Environment, Food & Rural Affairs
- Our World in Data — Oxford lifecycle analyses
- IPCC AR6 — Intergovernmental Panel on Climate Change
| Metric | Value |
|---|---|
| Food items tracked | 80+ with real CO₂e data |
| Transport modes | 15 from walking to long-haul flights |
| Energy sources | 9 including Quebec hydro (0.002 kg/kWh!) |
| Country benchmarks | 10 including Paris 1.5°C target |
| API cost | $0 for full functionality |
- 🔥 Firebase/Supabase persistent storage
- 📧 Weekly/monthly progress reports via email
- 📱 Barcode scanning for packaged products
- 👨👩👧👦 Household mode (track family footprint)
- 🌱 Carbon offset marketplace integration
- 📲 Mobile PWA wrapper
- 🏆 Social features — compare with friends, community challenges
Khalil Limem
Full-stack developer passionate about using AI to solve real-world problems.
Built with 🌍 for the Frostbyte Hackathon 2026