GeoPrompt

Inspiration

Across Africa and other emerging regions, land is one of the most underutilized assets due to fragmented data, outdated systems, and a lack of accessible tools. We were inspired by the idea that anyone should be able to find the best land for any purpose, just by asking a question. The way people use ChatGPT to write emails or solve problems? We want them to use GeoPrompt to find where to build, farm, invest, conserve, or plan, without needing to be a GIS expert. This was our hackathon challenge: democratizing geospatial intelligence.

What it does

GeoPrompt is a geospatial AI land suitability search engine. You can type a simple prompt like: “Where should I build a solar farm in Lesotho?” “Show me top land for bamboo that’s flood-resilient.” “Where is the best land to start affordable housing outside Maseru?” GeoPrompt processes the query using natural language understanding, combines it with satellite data, terrain, soil, zoning, and climate layers, then returns: A dynamic map with suitability scores Key data (elevation, water access, zoning risk, climate resilience) Downloadable reports and API outputs It brings Google Earth + ChatGPT + Land Suitability together in one interface.

How we built it

I had chatgpt help me device a bolt.new prompt and i kept iterating from there.

Challenges we ran into

Handling large raster datasets with minimal latency was tough in a hackathon timeframe. The Bolt.new platform kept freezing a lot of times.

Accomplishments that we're proud of

Built an end-to-end AI-driven system that lets users discover high-potential land from a simple prompt. Created an inclusive interface that removes the need for GIS expertise. Integrated multiple satellite and environmental data sources in a unified model. Used real data from Lesotho and Southern Africa for testing, showing its potential for scale. Recognized how transformative this can be for governments, small businesses, and communities.

What we learned

I learned that I can really do anything if I put my mind to it. I built this MVP in less than 5 hours. Also, as a non-technical person, this is my first ever "vibe coded" project, I learned how to fix bugs and understand the flow of building a project.

What's next for GeoPrompt

Build a mobile-first version for users in rural areas and low-bandwidth regions. Launch a public beta in Lesotho with local planners, students, and entrepreneurs. Expand coverage to all 10 districts and begin onboarding other African countries. Partner with ministries, urban planners, climate funders, and startups needing site intelligence. Introduce a "GeoPrompt Studio" API for developers to build apps on top of our engine.

Built With

• lucidereact
• react18
• tailwindcss
• typescript
• vite