TerraSense
An AI assistant for farmers that plans the growing season, spots crop disease from a photo and tracks harvests. It won 1st place at the 2025 Emeris Hackathon.
The code is private for now
Tech stack
- Next.js
- Tailwind CSS
- Supabase
- PostgreSQL
- Supabase Edge Functions
- Google Calendar API
- Weather and soil data APIs
- AI image recognition
My part
Built by a team of four at the 2025 Emeris Hackathon, where it won 1st place. I co-authored the research paper based on it, which is under review for SATNAC 2026.
What it does
- An AI crop calendar that plans planting, fertilising and harvesting from live weather, soil type and the season
- Syncs that calendar and its tasks to the farmer's own Google Calendar
- Upload a photo of a plant to detect disease, nutrient deficiencies or pests, with treatment recommendations
- A harvest tracker that logs yield by crop, field and season, with the conditions at the time
- A marketplace where farmers list their produce, searchable by location and transport needs
How it works
- Supabase provides the database, sign-in, file storage and server functions, so the team could focus on the farming features.
- Planning and image recognition come from external AI models, combined with weather and soil data.
Planned next
- A live feed of local and global commodity prices
- Simple price trend analytics to help farmers decide when to sell