SeeMore AI: What If AI Made Us Look Up Instead of Stare at Screens?

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

🌿 SeeMore AI — Look Closer. See More.

What if AI didn't give you the answer, but gave you a reason to discover it yourself?

We look at the world every day.

We walk past trees without noticing their patterns. We sit beside windows without observing their designs. We see everyday objects so often that we stop paying attention to them.

But what if we could turn these ordinary moments into opportunities to observe, think, and discover something new?

That question inspired me to build SeeMore AI.

👀 What I Built

SeeMore AI is an AI-powered observation training tool that transforms photos of everyday surroundings into interactive observation challenges.

The idea is simple: instead of directly identifying objects in a photograph, AI creates an indirect clue that encourages the user to observe and figure out the answer.

Imagine uploading a photograph containing a house, a tree, windows, and a streetlight.

Instead of saying:

“Find the tree.”

SeeMore could challenge you with:

“Find something in the scene that provides natural shade for pedestrians.”

The user must connect the clue with the environment and identify the answer independently.

AI gives the clue. Human does the observing.

The goal is to encourage attention to detail, pattern recognition, and visual reasoning through everyday surroundings.

SeeMore is designed for students, curious learners, and anyone who wants to become more observant.

💡 Where Did the Idea Come From?

When we think about AI, we often think about getting answers faster, automating tasks, and spending more time interacting with technology.

I wanted to explore a different possibility.

What if we used AI to encourage people to interact more with the world outside their screens?

I started imagining an application where a user uploads a photograph, receives a clue, and then has to look carefully at the real environment to solve it.

The AI would not simply point to an object or reveal the answer immediately. It would create a small challenge that makes the user think.

That became the central idea behind SeeMore AI.

I chose this approach because observation is a skill we can practise in ordinary places. A classroom, garden, street, or even our own home can become an opportunity to notice details we usually overlook.

And that connects directly with the Touch Grass challenge.

I didn't want to build another experience that simply demands more screen time. I wanted to experiment with using technology as a starting point for real-world observation.

🌱 How SeeMore Connects With Touch Grass

SeeMore follows a different interaction pattern:

AI gives a clue → You look around → You observe → You think → You answer.

The screen is only the starting point. The user is encouraged to pay attention to the real world rather than relying entirely on the AI to do the thinking.

For example, a clue about natural shade could encourage someone to notice trees around their neighbourhood. A clue about repeating patterns could help them discover details in windows, railings, or buildings.

The intention is to make familiar surroundings feel interesting again.

Of course, the experience depends on the user actually observing their environment. SeeMore is designed to encourage that behaviour rather than claiming that using an application automatically improves attention.

The goal isn't to keep people staring at a screen. It's to give them a reason to look up.

⚙️ How It Works

The intended experience follows these steps:

1. Upload a photo

Choose a photograph of everyday surroundings.

1. Analyse the scene

The AI examines visual information and looks for useful objects, relationships, patterns, or functions.

1. Receive an indirect clue

Instead of immediately revealing the target, the system presents an observation challenge.

1. Observe and think

The user studies the scene and works out the answer.

1. Submit the answer

The user enters their observation.

1. Receive feedback and a score

The application is designed to evaluate the response and provide an observation score, with the aim of making each challenge an opportunity to learn.

The project also includes a dashboard concept for tracking progress and encouraging users to continue practising.

🛠️ How I Built It

I built SeeMore AI as a full-stack web application using AI-assisted development.

Here are the main technologies used in the project:

Google AI Studio: The environment I used to build and develop the application with AI assistance.
Gemma Open Models: The intended core AI model for image understanding and generating indirect observation clues.
React: For the interactive frontend.
TypeScript: For application logic and type safety.
Vite: For frontend development and building.
Node.js and Express: For the backend and API layer.
Tailwind CSS: For the user interface styling.
Google Gen AI SDK (@google/genai): For integrating the Google AI model API.
GitHub: For source code management and sharing the project.

I designed the interface around a dark visual theme with cyan and violet accents, keeping the focus on the observation experience.

One important design decision was to make the clue indirect. If the AI immediately reveals the target, the user loses the opportunity to observe and reason independently.

The challenge is not simply about identifying an object. It is about making the user think about what they see.

Note: The model name and runtime configuration should match the actual working configuration of the deployed application.

🔓 Why Does Open Innovation Matter?

For SeeMore AI, the AI model is central to the experience.

An open-model approach creates opportunities to experiment with how visual information is interpreted and how observation challenges are generated.

It also offers flexibility for future development. Different models and deployment approaches could be explored without having to redesign the entire user experience around one particular provider.

In future versions, I would like to investigate model customisation, different difficulty levels, educational use cases, and privacy-conscious deployment options.

Open innovation matters because it gives developers room to learn, experiment, adapt, and build on shared technology.

For a project like SeeMore, that flexibility could help make observation challenges more useful for different users and environments.

I also recognise that using an open model does not automatically make an application offline or guarantee that user data never reaches a server. Those capabilities depend on the actual model deployment and application architecture.

🚀 What's Next for SeeMore AI?

This project began as a Hacktoberfest challenge, but I would love to develop it further.

Some possibilities I want to explore include:

• More diverse observation challenges.
• Better difficulty adaptation.
• More detailed progress tracking.
• Outdoor exploration activities.
• Educational modes for students.
• Improved answer evaluation.
• Additional open-model deployment options. My long-term goal is to turn everyday environments into opportunities for curiosity and learning.

Because sometimes, we don't need a new place to explore.

We just need to learn how to look at the world we're already in. 🌿

🎥 Demo

Live Demo: https://seemore-aibyshara.ai.studio

Explore SeeMore AI and discover how an ordinary photograph can inspire an observation challenge.

💻 Code

GitHub Repository: https://github.com/shraddhakolate30-maker/seemore-ai

The repository contains the application's source code, project structure, configuration example, and documentation.