🌱 Touch Grass: TrailLens – Offline AI Nature Identifier
For Hacktoberfest 2026, I’m building TrailLens, an AI-powered tool that encourages people to spend more time outdoors by helping them identify plants and leaves while hiking, walking, or exploring nature.
📸 What it does
- Take a photo of a plant or leaf outdoors
- Use an open-weight AI model to identify it
- Provide basic information about the identified plant
- Designed to work locally, reducing the need to upload images to a remote server
- Makes the screen useful for only a few seconds — then gets you back outside 🌿
🤖 Why open-source AI?
The open approach is an important part of this project. Using an open-weight model gives us the ability to:
- Run AI locally
- Keep personal photos on the user's device
- Experiment with and modify the model
- Swap models without being locked into one provider
- Reduce dependence on paid cloud APIs
- Potentially use the application in places with poor or no internet connectivity
This is where open innovation works better for this project than a closed, cloud-only approach: the AI can come to the trail instead of requiring the trail to have an internet connection.
🌍 Why “Touch Grass”?
AI shouldn't always keep us in front of a screen. I wanted to build something where technology is simply a bridge to the real world.
The goal is simple:
Take your phone → identify something → learn about it → put your phone away → explore. 🌱
🛠️ Tech
- Python
- Computer Vision
- Open-weight AI model
- Local inference
- Image Processing
This project is still evolving, and I’m excited to experiment with open-source AI and take it outside!
Top comments (0)