I Built a Yard Care App with AI: A Personal Experiment
A journalist used Gemini AI to create a custom Android app for managing his yard, facing both successes and challenges along the way.
When I returned to my computer five minutes after giving Gemini a lengthy prompt, I had two things: a functional app in a preview window, and a message about a bug. Below it was a button to fix it. I clicked, and after 233 seconds Gemini reported success, using terms like "blockages" and "race conditions" that I didn't understand. It was thrilling.
This was my second or third attempt at vibe-coding. The previous project was a web app with one job: check if a local high-end grocery chain is running its annual Peach-o-Rama event. No peaches yet. This time, the goal was more ambitious: an app to help me master my unruly yard.
When my husband and I moved into our house eight years ago, we didn't think much about yard work. We ignored it until weeds took over. Flower beds filled with biblical proportions of weeds. We called a landscaper whose one-time visit put the yard on autopilot for years. It worked, but then weeds returned and shrubs showed distress. This spring, I resolved to figure out what was going on.
I crafted a detailed prompt for Gemini: manage a long list of yard care chores, make recommendations, take weather into account, use image recognition for plant diagnosis. I entered it into Google AI Studio to create an Android app. Within minutes, I had a working preview with logical sections for different plant zones and an AI "plant doctor" where I could upload images.
But there was a color scheme issue: Gemini chose dark mode with dark purple and brick red accents, making text illegible. I suggested a white background with light green, pink, and blue. Gemini returned with something nicer and a greeting: "Welcome Back, Gardeneer!" I kept that.
I kept the basic structure but requested tweaks, like integrating live weather data instead of climate presets Gemini had created. I uploaded the app to my phone and started using it, though I immediately noticed issues: I couldn't edit chores once created, schedule them for specific days, or link plant profiles to tasks. The date picker didn't work.
On the other hand, the AI plant doctor worked great out of the box. I uploaded a picture of a sick rhododendron, and after a minute it gave a detailed health report (critically bad!), likely factors, and action items. That was exactly the help I needed.
It turned out the landscaper's fix—covering beds with landscape fabric and river rock—was problematic. Gemini said the fabric and rocks were suffocating the root system and the sun-baked rocks were cooking the roots. No wonder my yard looked bad.
The next day, I followed Gemini's recommendations: removed rocks and fabric from the rhododendron, pruned branches. After a few days, new leaves appeared. Yard work turned out to be extremely satisfying. I learned that AI doesn't know the real world—it doesn't grasp legibility or real-time weather. But building the app was instructive. While my "Gardeneering" app may never reach the Play Store, it taught me the importance of clearly defining the problem before coding.
