-
-
Design your garden your way! List your plants and plot size and build a garden from the generated graph.
Inspiration
Going through the pain of designing a garden concept from scratch with my wife, both of us being total noobs, was excruciating. In reflecting on ways to really improve the quality of people's lives in today's world, I realized that agentic platform could give me the tools to make the exact solution I wanted. Better yet, this solution could benefit the masses by lowering the bar to entry for starting a garden, or prototyping new arrangements as a master. Thus, the generative garden plot designer app was born and it was time to Flourish!
What it does
Flourish is a native application for Android devices in which allows the user to input three key pieces of data: Their Location, Their Plot Size, and What Plants They Want. It guides the process by monitoring for compatible grow zones and making the selection and easy to tinker experience. Once the selection is made, one button and an interactive garden plot diagram is made followed up with critical data for understanding and maintaining the plant health. At the end, the user can choose to save their journey as a PDF instruction, or they can reset and do it all over again in minutes instead of hours.
How we built it
Using Claude AI, database files were aggregated to create a list of almost 100 edible plants with key data about their properties, behavior, companions, and pest control. An additional database was then generated to represent all of the grow zones across the Unites States of America. These two files were ported directly into Android Studio where the application was made through agentic and manual coding. Legibility and practical functionality was prioritized to MVP a product that deploys natively on almost any Android phone and gives the user unlimited access to making a garden adventure guide.
Challenges we ran into
Data selection, feature creep, and agentic spasms caused moments of uncertainty in the quality and capabilities of achieving the goals. Through detailed code review and structured resets of the agents, a micro step approach to developing was made possible within the timeframe achieving around 95% of the MVP features desired. There was also a learning curve for emulating the phone virtual machine for testing as a moment came when the whole system crashed and I had to dive into the failure modes to recover and continue testing and development.
Accomplishments that we're proud of
The application is fully functional and tested to be modular to varying phone sizes and scalable garden designs. The data provided to the user is practical, detailed, but not overly complex so anyone can pick it up and dive right into preparing the freshest of feasts. Edge case detection to simplify the user experience was implemented to make the product close to production quality and very user friendly.
What we learned
Gardening is a real passion for people. Collaborating with agentic AI models is really practical and can turn out phenomenal results in record time. I was really hesitant to trust the code quality and commenting content at first, but I learned to lean into the abyss and was able to see the workflow in a new light.
What's next for Flourish: Generative Garden Design App
My vision for Flourish is to keep expanding on the base feature design through more voice of the customer extraction. Once a depth of experience is developed, the path pivots towards an app that grows with the gardener, develops dynamic recommendations with balanced plot maintenance, allows for more than edible plant architecture, connects people with local food coops, and directly connects the customers with the suppliers who can bring their concept straight to their door for a fully functional garden in their living space.
Log in or sign up for Devpost to join the conversation.