Build a garden plan with local AI

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

What I Built

I built Sprout Scout, a local-first garden planner that helps people decide what to plant, when to plant it, and what small outdoor task to do next. The app asks for a region (optional), the estimated last spring frost date, and the plants the user wants to grow, then generates a simple planting window and a short outdoor plan.

The goal is to make the screen the shortest part of the experience. Instead of spending time searching online, the user gets a quick plan and then goes outside to check sunlight, prepare a patch, or start a small gardening task. It’s built for beginners, home gardeners, and anyone who wants a gentle push to spend more time outdoors.

Demo

This project runs locally with Ollama and an open-weight model.

powershell
ollama pull llama3.2:latest
python server.py

CODE : https://github.com/deepanshu-singh22/local-garden-planner/tree/main

How I Built It
I used the open-weight model llama3.2:latest through Ollama for local inference. The app uses a small Python backend to call the local model, validate the output, and return a structured garden plan to the browser. The frontend is built with plain HTML, CSS, and JavaScript, so the whole project stays lightweight and easy to run on a laptop.

The open-source AI stack is the core of the product: it keeps the experience local, private, and flexible. If a different model is needed, it can be swapped without changing the app’s overall workflow.


Why Does Open Innovation Matter?
Open innovation matters here because the app is designed to work on the user’s own machine instead of sending garden details to a cloud AI service. That means:

no API subscription
no per-request charging
local privacy
works without internet after setup
easy model swapping and customization
A closed AI API could have created a prototype faster, but it would have introduced cloud dependence, hidden costs, and less control over what the model does. With open-weight tooling, I could build a practical tool that keeps user data local and makes it easy to adapt the project for different gardening needs.


![ ](https://dev-to-uploads.s3.us-east-2.amazonaws.com/uploads/articles/y4p4zf7z6jn2r9bif8cw.png)