🌱 GardenBuddy: An AI Garden Agent That Gets You Off the Screen

*This is a submission for the [Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass]

What if an AI application helped you close your laptop instead of keeping you chatting?

That was the idea behind GardenBuddy, my project for Hacktoberfest 2026 Week 1: Touch Grass.

Instead of building another general-purpose chatbot, I wanted to build something practical: an AI gardening assistant that helps users organize their gardens, consider weather conditions, and get a plan for what to do outside.

🔗 Try GardenBuddy · GitHub Repository

🌿 What I Built

GardenBuddy is an AI Garden Weather Agent with a simple goal:

Help people spend less time on screens and more time growing things.

Here's what the application looks like.

The dashboard displays the daily garden plan, weather status, selected garden, and plant-management section.

The main features include:

• 🌱 Create a garden with its location, type, and sunlight exposure.
• 🪴 Add plants with species, planting dates, locations, and care notes.
• 🌦️ Use weather information to inform gardening plans when the weather service is available.
• ✅ Review suggested tasks and use them to guide real gardening activities.
• 🧠 Build toward more personalized planning using AI reasoning and persistent memory.

🪴 How to Use GardenBuddy

1. Create Your Garden

Start by clicking + New in the top navigation bar.

Enter the following information:

The location is especially important because weather and growing conditions vary by region.

Click Save Garden to create your garden.

2. Add Your Plants

Once your garden is created, click + Add Plant.

For example, you could add a tomato plant:

Click Save Plant.

You can repeat this for coriander, spinach, mint, or any other plants you're growing.

The purpose of recording this information is to give GardenBuddy useful context about the garden rather than treating every request as an unrelated question.

3. Get Your Daily Garden Plan

After selecting your garden, open Today's Daily Garden Plan and use Refresh Plan when you want to generate or refresh the plan.

A plan might include tasks such as:

• Check soil moisture before watering.
• Prepare a garden bed.
• Review which crops are suitable for your growing conditions.

Weather-aware recommendations depend on the weather service returning valid data. If the service is unavailable, the application should communicate that limitation rather than pretend it has a reliable forecast.

The intended workflow is straightforward:

Create garden
     ↓
Add plants
     ↓
Get daily plan
     ↓
🌱 Go outside
     ↓
Complete garden tasks

That's the part of the project I care about most: turning an AI-generated plan into a real-world activity.

🏗️ How I Built It

I structured GardenBuddy around separate responsibilities instead of putting everything into one large AI prompt.

Here's the intended architecture:

              🌱 GardenBuddy
                    │
              Frontend (Vercel)
                    │
                    ▼
                 FastAPI
                    │
          ┌─────────┼──────────┐
          ▼         ▼          ▼
     PostgreSQL  Weather API  AI Layer
                               │
                     ┌─────────┴────────┐
                     ▼                  ▼
               Thinking Machines    Backboard.io
                 AI reasoning       AI memory

              RAG → Gardening knowledge

              Render → Backend hosting

Technology choices

DevRelay MCP is part of my exploration of agent tooling. MCP can expose tools to an AI agent, but it does not provide weather data by itself.

I want each technology to solve a real problem rather than add tools just to make the stack look impressive.

Project structure

The project is organized around the frontend, backend, and project documentation.

garden_buddy/
│
├── PROJECT.md       # Product requirements and architecture
├── AGENTS.md        # Coding-agent instructions
├── TASKS.md         # Development roadmap
├── SECURITY.md      # Security boundaries
├── README.md        # Project documentation
│
├── frontend/        # Garden dashboard and forms
├── backend/         # FastAPI application
│
└── .env.example     # Required configuration names

This illustrates the intended organization; check the repository for the exact current directory and file names.

I also defined project instructions for AI-assisted development. They explain what the application should do, how implementation should proceed, and which security boundaries must be respected.

How the components fit together

The core workflow is:

Garden + Plant Data
        │
        ▼
Weather Information
        │
        ▼
Relevant Gardening Knowledge
        │
        ▼
AI Reasoning + Garden Context
        │
        ▼
Today's Garden Plan
        │
        ▼
Real-world Gardening 🌱

The important distinction is that the AI should use available garden information and actual weather data. It should not invent a forecast or assume that a plant needs watering without considering the situation.

💻 Code and Demo

🌐 Live demo: https://gardenbuddy-frontend.vercel.app/

🐙 Source code: https://github.com/PrabodhKumar01/garden_buddy

I'd welcome feedback, bug reports, and contributions that make GardenBuddy more useful for real gardeners.

🌍 Why Does Open Innovation Matter?

For me, open innovation is about being able to inspect, experiment with, and improve the technology behind an application.

With an open-source project like GardenBuddy, contributors can help improve the garden-planning workflow, add regional gardening knowledge, test recommendations, and explore different AI approaches.

It also makes the project easier to extend without committing its entire future to one provider.

The broader opportunity is to build AI that supports useful activities outside the digital world.

🏆 Prize Categories

These are the partner tracks relevant to the technologies I am exploring. Actual eligibility depends on the challenge rules and the integrations demonstrated in the submitted version.

• 🚀 Best Use of Render: Hosting the GardenBuddy backend and supporting its deployment workflow.
• 🧠 Backboard.io: Persistent garden memory, where implemented and demonstrated.
• 🤖 Thinking Machines: AI-powered garden planning and reasoning, where implemented and demonstrated.

Each integration should have a clear responsibility and a working demonstration. I don't want to claim that a planned integration is already complete.

🌱 What's Next?

My next priorities are to improve weather reliability, strengthen personalized recommendations, expand the gardening knowledge base, and test the application through actual gardening activities.

I'd also like to improve how GardenBuddy remembers observations over time so that future plans can make better use of a garden's history.

Final Thoughts

Hacktoberfest's Touch Grass theme challenged me to think beyond building another application that keeps people online.

GardenBuddy is my attempt to use AI for something simple and tangible: helping people care for plants and spend more time outdoors.

The idea is straightforward:

Let AI help with the planning. Let people do the gardening. 🌱

Thanks for reading! If you have ideas for improving GardenBuddy, I'd love to hear them.