This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass.
🌱 GardenMitra — Your Green Friend
Grow something real. Touch grass.
What I Built
GardenMitra is an open-source garden companion that uses Gemma AI running locally through Ollama to help people understand their plants, plan growing spaces, and take practical gardening actions.
We spend hours searching for answers on our screens, even when what we really want is to get outside and do something meaningful.
Gardening can be intimidating for beginners. Which plants suit a balcony? Why might a plant be struggling? How should a growing space be organized? GardenMitra aims to make those decisions easier and turn gardening into a habit.
Whether you have a few pots on an apartment balcony, a terrace garden, or an indoor plant collection, GardenMitra is designed to make plant care more approachable.
🌿 What GardenMitra offers
- AI gardening chat: Ask gardening questions and get assistance from Gemma running locally through Ollama.
- Plant image analysis: Submit a plant or garden image for AI-assisted visual analysis.
- Garden Doctor: Get structured guidance for common plant-care concerns.
- Garden planning: Organize growing spaces and garden zones.
- Sunlight intelligence: Use available observations and calculations to reason about growing conditions.
- Care tasks: Break plant care into manageable actions.
- Shopping list: Keep track of supplies needed for upcoming gardening activities.
GardenMitra combines AI assistance with deterministic gardening workflows. Not every recommendation or calculation is generated by an AI model.
The goal is to help users move from uncertainty to action: understand what needs attention, put the phone away, and spend time caring for their plants.
🌎 Why Touch Grass?
Technology should help us experience more of the world, not keep us looking at screens all day.
GardenMitra encourages a simple routine:
- Check your garden and identify something that needs attention.
- Ask for guidance when you need it.
- Choose one manageable gardening task.
- Step away from the screen and do it.
- Return when you need to plan or record progress.
The best gardening session is not the one spent chatting with AI. It is the one that helps you grow something real.
Demo
- Live application: https://gardenmitra.ai.studio/
- GitHub repository: https://github.com/DEVDESAGAR1/gardenmitra
- GitHub code scanning: https://github.com/DEVDESAGAR1/gardenmitra/security/code-scanning
Code
Repository: GardenMitra on GitHub
GardenMitra is being developed as a modular gardening application with a focus on practical workflows, understandable logic, and extensibility.
The repository is the primary reference for the implementation and setup instructions needed to run the application and reproduce the AI workflows.
How I Built It
I wanted to build something more useful than a generic gardening chatbot: an application that combines conversational assistance, visual input, and practical garden-management workflows.
🧠 Local AI with Gemma and Ollama
GardenMitra uses Gemma through Ollama for local AI inference.
The two AI workflows are:
- Gardening chat: The user asks a gardening question and receives a response from the locally running model.
- Image analysis: The user supplies a plant or garden image for AI-assisted analysis.
Running inference locally allows the AI to operate without sending every prompt to a hosted model provider, provided the application uses the local Ollama endpoint and no other component uploads that information.
The exact Gemma model version matters, particularly for image analysis. The repository should identify the model tag and document its supported capabilities.
🏗️ Engineering approach
The project combines AI-assisted workflows with conventional application logic:
- Local Gemma inference through Ollama.
- Text-based gardening assistance.
- Image input for AI-assisted analysis.
- Garden-space and zone organization.
- Structured, rule-based care guidance.
- Sunlight calculations where supported by available inputs.
- Browser local storage for the initial experience.
- Automated checks and production-build verification during development.
Weather data is simulated by the mock weather provider; GardenMitra does not claim that this provider supplies live weather information.
AI-generated plant guidance and image analysis should be treated as assistance, not a guaranteed diagnosis. Results can be incomplete or incorrect, so users should verify important plant-care decisions.
Reproducibility
- AI runtime: Ollama.
- Model family: Gemma.
These details help other developers reproduce the setup and understand which model capabilities GardenMitra uses.
Why Does Open Innovation Matter?
Gardening knowledge should be accessible, adaptable, and useful to people with different growing spaces, budgets, climates, and levels of experience.
Using open-weight AI locally gives GardenMitra an alternative to making every interaction dependent on a proprietary cloud API.
Privacy and control
With a local Ollama setup, prompts and images can be processed on the user's own machine instead of being sent to a hosted model provider. This depends on the actual deployment: other application services must also avoid uploading that information if the entire workflow is to remain local.
Accessible experimentation
An open-weight model gives developers the opportunity to experiment with different configurations and adapt the application to their needs.
Local inference can also avoid per-request charges from a hosted inference API, although it still requires suitable hardware and consumes computing resources.
Community-driven gardening
A balcony gardener in Mumbai may need different guidance from someone growing plants in a colder climate. An open-source application gives developers and gardening enthusiasts room to improve plant knowledge, refine workflows, and contribute region-specific improvements.
Open innovation is about more than choosing a model. It is about making the technology behind a useful tool easier to experiment with, adapt, and improve.
My principle is simple: technology should help people care for the real world, not replace their experience of it.
Prize Categories
- Hacktoberfest Open-Source AI Challenge Week 1 — Touch Grass
- Best Use of Gemma — intended category, subject to the official category rules and verification of the working implementation.
Final Note
GardenMitra brings together local AI, practical gardening workflows, and the belief that technology should encourage real-world action.
I built it to make gardening less intimidating and encourage people to spend more time growing, observing, and caring for plants.
Built by Sagar Devde.
🌱 GardenMitra — Grow something real. Touch grass.
Top comments (0)