🍃 What I Built
Touch Grass AI is an open-source, offline-first outdoor companion designed to help people reconnect with nature while reducing unnecessary screen time.
Whether you're hiking on remote trails without cellular service, identifying plants in your local park, or recording bird calls in your backyard, Touch Grass AI aims to bring nature exploration and AI-powered identification together in one accessible platform.
The goal is simple: use technology to help people spend more time outdoors, not more time on their screens.
👥 Who Is It For?
- 🥾 Hikers & Trail Runners: Explore outdoor routes, discover nature, and generate GPX routes for offline use.
- 🌱 Gardeners & Foragers: Identify plant species, maintain species logs, and explore local frost and microclimate information.
- 🐦 Birdwatchers & Nature Enthusiasts: Analyze audio spectrograms and explore AI-powered bird-call recognition.
⏱️ Designed to Reduce Screen Time
Unlike traditional apps designed to maximize engagement, Touch Grass AI encourages short, purposeful interactions.
Its 15-Minute Outdoor Micro-Quest feature encourages users to complete a nature-based activity, put their phones away, and engage with the environment around them.
The idea is to make technology a companion to outdoor exploration rather than a distraction from it.
🚀 Live Demo
🌐 Live Application:
https://hacktoberfest-week1.onrender.com
🩺 API Health Endpoint:
https://hacktoberfest-week1.onrender.com/api/health
💻 GitHub Repository:
https://github.com/PruthviRajG25/Hacktoberfest-Week1
⚡ Test the API
You can check the deployed backend directly from your terminal using the following command:
curl -X GET https://hacktoberfest-week1.onrender.com/api/health
The expected response is:
{
"status": "online",
"appName": "Touch Grass AI",
"openAIWeights": "Pre-Warmed Ultra-Fast ViT & AST Neural Engine (<15ms)",
"offlineModeSupported": true,
"timestamp": "2026-10-10T08:00:00.000Z"
}
Note: This is the expected response format provided for the project. Actual values and service availability may vary.
🏗️ How I Built It
Touch Grass AI combines a web-based outdoor experience with AI-assisted image and audio classification, interactive mapping, and local data storage.
The application is organized into two main parts:
1. Client-side experience
- 📸 Image capture and plant-species identification.
- 🎙️ Audio recording and spectrogram-based analysis.
- 🗺️ Interactive maps and outdoor route exploration.
- 💾 Local storage for an offline-friendly species journal.
2. Backend inference and APIs
- ⚡ Express.js REST API.
- 🩺 API health monitoring.
- 🌿 Image classification endpoint.
- 🎵 Audio classification endpoint.
🔄 System Architecture
flowchart TD
A["👤 Explorer / Hiker"] --> B["📱 Web Client / PWA"]
subgraph Frontend["Client-Side Features"]
B --> C["🗺️ Leaflet Route Builder"]
B --> D["🎙️ Audio Recorder"]
B --> E["💾 Offline Species Journal"]
end
subgraph Backend["Express.js Backend"]
B -->|REST API| F["⚡ Express Server"]
F --> G["🟢 GET /api/health"]
F --> H["📸 POST /api/classify-image"]
F --> I["🎵 POST /api/classify-audio"]
H --> J["🌿 Vision Classification"]
I --> K["🐦 Audio Classification"]
end🛠️ Technology Stack & AI Models
🌿 Vision Identification
Vision Transformer (ViT) and MobileNetV3 are the proposed vision models for classifying plant-related images, including leaves and trees.
The objective is to make species identification accessible without depending entirely on proprietary cloud AI services.
🐦 Audio Identification
The Audio Spectrogram Transformer (AST) is used in the planned audio-classification pipeline for analyzing recorded sounds and supporting bird-call recognition.
🗺️ Interactive Maps
- Leaflet.js: Interactive map rendering and outdoor route visualization.
- OpenStreetMap: Open map data for exploring outdoor locations.
- GPX export: Designed to support saving routes for use with compatible navigation applications.
⚙️ Backend & Deployment
- Node.js: JavaScript runtime for the backend.
- Express.js: REST API and backend routing.
- Render: Application hosting and deployment.
- Local storage and offline-friendly web capabilities: Intended to keep supported journal data accessible without a continuous internet connection.
🌐 Why Does Open Innovation Matter?
For outdoor applications, open innovation can make a meaningful difference in accessibility, privacy, and reliability.
1. 📶 Reliability Beyond Connectivity
Remote trails and forests often have limited network coverage. Local inference and offline-capable features can help users continue supported activities even when connectivity disappears.
2. 🔒 Privacy by Design
Processing supported photos, recordings, and location data locally can reduce the need to upload sensitive information to third-party services.
The long-term goal is to give users more control over their outdoor data and nature observations.
3. 💸 Lower Dependence on Paid APIs
Open-weight models can reduce reliance on metered AI services and recurring per-request charges. This creates opportunities to build accessible tools without making every identification dependent on a paid cloud API.
4. 🌍 Community-Driven Development
Open-source development allows contributors to improve classification, expand species coverage, refine offline support, and build features that benefit nature enthusiasts everywhere.
🤖 My AI Agent Development Workflow
AI-assisted development workflows supported the project's implementation, testing, and deployment process.
DevRelay CLI integration: The development session was logged and managed through DevRelay CLI tools.
The project also provides an opportunity to explore how AI-assisted engineering workflows can help developers prototype and iterate on practical, open-source applications.
🏆 Hacktoberfest Challenge Categories
- 🌿 Main Challenge: Hacktoberfest Week 1 — Touch Grass
- 🤖 Partner Category: Open-Weight AI
- 📱 Partner Category: Local / Offline Inference
🌱 What's Next?
Touch Grass AI is built around a simple idea: technology should help us discover more of the world beyond our screens.
Future improvements can focus on expanding species coverage, improving model performance, strengthening offline functionality, and making outdoor exploration more accessible.
If you're interested in open-source AI, nature exploration, offline-first applications, or contributing to projects that encourage healthier technology habits, check out the repository.
🔗 Explore the code: https://github.com/PruthviRajG25/Hacktoberfest-Week1
🚀Try the application: https://hacktoberfest-week1.onrender.com
🌿 Get outside, explore your surroundings, and go touch grass!
Top comments (0)