What If AI Helped Us Notice the World Around Us? 🌿
We use our phones to learn, work, communicate, and even relax. But sometimes, we get so used to looking at a screen that we stop noticing what's happening around us.
The sound of birds outside. The wind moving through the trees. A small plant we've walked past a hundred times without really looking at it.
I started thinking about this while working on the Hacktoberfest Open-Source AI Challenge — Week 1: Touch Grass.
Instead of building another AI tool that asks people to spend more time on their devices, I wanted to try something different.
That's how MemoryWalk started.
What I Built
MemoryWalk is a small web application powered by locally running AI. It gives you an observation mission to complete in the real world, encourages you to put your phone away, and lets you reflect on what you noticed when you return.
The idea is simple: AI gives you something to notice. Then you put the phone away.
Here's how it works:
- Choose how much time you have and what you want to focus on.
- Get an AI-generated observation mission.
- Put your phone away and explore your surroundings.
- Return and answer a few reflection questions.
- Receive AI-generated feedback, a score, and a personal takeaway.
- Save the experience in your Memory Garden.
For example, a mission might ask you to become a sound detective for ten minutes. Instead of listening to music or scrolling, you pay attention to birds, passing vehicles, the wind, and sounds you usually ignore.
Another mission might ask you to observe the colors and patterns of plants around you.
These are ordinary things, but the goal is to experience them more intentionally.
Demo
The best way to understand MemoryWalk is to see the experience from beginning to end.
The demo should show the home screen, generating a mission, completing the reflection flow, and saving a memory in the Memory Garden.
🎥 Demo video:
💻 GitHub repository:
🌿 MemoryWalk — Look Up. Look Around.
AI gives you something to notice. Then you put the phone away.
MemoryWalk is a local-AI-powered experience designed to help people step away from their screens and pay attention to the world around them.
Choose how much time you have and what you want to notice. MemoryWalk generates a small observation mission using a locally running AI model. Put your phone away, go outside, observe the real world, and return to reflect on what you experienced. Turn that moment into a personal memory in your Memory Garden.
✨ What It Does
- Personalized observation missions: Generate a mission based on your available time and chosen focus.
- Explore the real world: Observe nature, listen to your surroundings, discover tiny details, or let curiosity choose.
- Phone-down experience: Encourages you to leave the screen behind while completing your mission.
- AI-powered reflection: Return and respond to reflection questions…
You can also run the project locally using the instructions below.
How I Built It
I kept the stack simple because I wanted to focus on the experience rather than setting up a complicated infrastructure.
The project uses:
- HTML, CSS, and JavaScript for the frontend.
- Python for the local web server and backend endpoints.
- Ollama to run the AI model locally.
-
qwen2.5-coder:3bfor generating observation missions, reflection questions, and feedback.
The browser communicates with the Python backend, which sends prompts to Ollama. The generated responses are then displayed in the application.
The basic architecture looks like this:
Browser → Python backend → Ollama → Local language model
One thing I wanted to explore was whether AI could be useful without becoming the main activity itself. In MemoryWalk, the AI starts the experience, but the important part happens away from the screen.
Building this also gave me an opportunity to work through the practical details of connecting a frontend to a local model, handling generated responses, and making the interface feel like a complete experience rather than just a chat window.
Why Does Open Innovation Matter?
For this project, open innovation matters because it gives developers room to experiment with AI in different ways.
Using Ollama with a locally runnable model allowed me to build and test MemoryWalk on my own computer without relying on a paid hosted-model API.
There are a few things I appreciate about this approach:
1. More control over experimentation
I can change prompts, test different missions, and explore other compatible models without redesigning the whole application.
2. A local-first approach
The model runs on my own machine, which makes local experimentation possible without sending every prompt to a hosted AI service. The application still needs to be assessed for its complete data handling and privacy behavior, but local inference gives me more control over where model processing happens.
3. Lower barriers for individual developers
You don't necessarily need a complex cloud setup to start experimenting with language models. A local runtime and a compatible model can be enough to build a useful prototype.
4. AI doesn't always need to keep us online
This is the part that connects most closely to the challenge. AI is often used to generate more content for us to consume. I wanted to experiment with using it as a starting point for something offline: observing, listening, reflecting, and remembering.
Open tools made it easier for me to explore that idea with the resources I had available.
What I Learned
One of my biggest takeaways was that the AI model is only one part of the experience.
The flow around it matters just as much. A generated mission needs to be clear enough to follow, the reflection should feel natural, and the application should encourage users to leave the screen instead of giving them another reason to stay.
I also learned to pay attention to the small implementation details that make a project usable, from connecting frontend files correctly to testing the full mission and reflection flow.
MemoryWalk is still a small project, and there is room to improve it. I'd like to explore more mission types, improve the Memory Garden, and make the experience even better on mobile devices.
But I like the idea behind it: technology can help us pay attention to life outside technology.
My Agent Session
I used AI assistance during development to help work through implementation questions, debug issues, and improve the interface. I tested the application locally as I built it and made adjustments when things didn't work as expected.
The project uses Ollama and qwen2.5-coder:3b for its local AI functionality.
What's Next?
I'd like to keep improving MemoryWalk and see how people use it.
What would you want your first mission to be: noticing nature, listening to the sounds around you, or finding tiny details you've never paid attention to before?
I'd love to hear your ideas in the comments.
Sometimes, the most interesting thing an AI can do is give us a reason to look away from it.
Look up. Look around. Keep the moments that matter. 🌿
Top comments (0)