This is my submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass.
github repo: https://github.com/saurabh18082005-cyber/TerraQuest-AI
🌿 What I Built
I built TerraQuest AI, an outdoor biodiversity explorer that turns an ordinary walk into a discovery mission.
The idea is simple: instead of spending more time scrolling, people can explore their surroundings, photograph plants and insects, and use AI-assisted identification to learn about nature.
The prototype includes:
- Time-based missions that encourage people to spend time outdoors.
- Photo-based discoveries for recording observations.
- AI-assisted identification that describes visible evidence in a photograph.
- Explorer profiles and leaderboard foundations for future community challenges.
- Responsible verification: AI suggestions remain unverified until independently confirmed.
The goal is to make exploring nature more educational, interactive, and fun—without disturbing wildlife.
🐞 Demo
I tested TerraQuest AI using a photograph of a ladybug.
The application successfully returned a discovery report identifying it as a ladybug and describing visible features such as red wing covers and black spots.
The report also clearly states that the identification is unverified and that the exact species requires further analysis.
[Insert your saved discovery screenshot here.]
💻 Code
GitHub repository: [Insert your public repository URL here.]
Tech stack: Python, Flask, HTML, CSS, JavaScript, SQLite, and OpenRouter for hosted vision-model inference.
🛠️ How I Built It
I started by building the core experience: generate an outdoor mission, upload a photograph, and display a discovery report.
The Flask backend receives the image and sends it to a vision-capable model through OpenRouter. It then processes the model's response and returns a report to the frontend.
The most challenging part was integrating the AI. I encountered model-access problems with earlier providers and had to troubleshoot inconsistent response formats before getting the discovery report working.
This taught me that building an AI application involves more than sending a prompt. Provider availability, response parsing, error handling, and honest uncertainty all matter.
🌍 Why Does Open Innovation Matter?
Open innovation gives developers the freedom to experiment with different models, learn from the community, and improve their applications without being locked into a single provider.
TerraQuest currently uses hosted inference. In the future, I want to explore open-weight models, stronger species verification, and local inference where practical.
I believe AI should help people learn about the world while making its limitations clear.
🧠 What I Learned
This project taught me to troubleshoot API integrations, handle model responses, protect credentials, and distinguish an AI suggestion from a verified scientific observation.
Seeing the discovery report finally appear after working through those problems was the most satisfying part of the build.
🚀 What's Next?
- Improve identification accuracy and verification.
- Add evidence-backed rarity scoring.
- Expand the shared leaderboard and discovery community.
- Explore open-weight vision models.
TerraQuest AI is still a prototype, but it is my first step toward making outdoor exploration feel like a real-world quest.
I'd love to hear what features you'd add to a biodiversity explorer!
Hacktoberfest #OpenSource #AI #Biodiversity
Top comments (0)