Guidely - Your indoor navigation partner

Inspiration

Getting lost is frustrating because maps often tell you where to go, but not what you're actually looking at. Traditional navigation relies on abstract maps, street names, and compass directions that can be difficult to follow, especially for first-time visitors, tourists, children, seniors, or people with cognitive disabilities.

We wanted to create a navigation experience that feels more natural, one that guides users using the world around them instead of a blue dot on a map.

What it does

Guidely is an AI-powered visual navigation assistant that helps users reach their destination using photos instead of traditional maps.

The user simply enters a destination and points their camera at their surroundings. Guidely analyzes what it sees, understands nearby landmarks, and overlays simple visual guidance directly onto the camera view.

Instead of saying:

"Walk north for 200 meters."

Guidely might say:

"Walk toward the red brick building."

or

"Turn right at the coffee shop."

To make navigation feel approachable, Guidely uses friendly illustrated characters and clear arrows that point users in the right direction.

How we built it

We built Guidely using Python for the backend, where AI analyzes images captured by the user and identifies landmarks in the environment.

Our prototype combines:

Computer vision to understand the scene A large language model to generate natural navigation instructions Mapping services to calculate routes Image annotation to overlay arrows, highlights, and friendly guide characters directly onto photos

The frontend provides an intuitive interface where users can enter a destination and receive easy-to-follow visual guidance.

Challenges we ran into

One of the biggest challenges was determining a user's location and orientation from a single image.

Unlike traditional GPS navigation, our approach requires understanding both the environment visible in the photo and how it relates to the user's destination.

Another challenge was balancing technical accuracy with simplicity. We wanted the interface to remain friendly and accessible while still providing useful navigation.

Accomplishments that we're proud of

Building a working proof of concept that combines AI vision with navigation.
Creating navigation that feels more intuitive than traditional maps.
Designing an interface that uses playful visual guides instead of overwhelming users with information.
Exploring a new interaction model where AI understands the user's surroundings rather than simply displaying a route. ## What we learned This project taught us that navigation isn't just about finding the shortest path, it's about making directions easy to understand.

We also learned how computer vision, mapping APIs, and modern language models can work together to create more human-centered experiences.

Most importantly, we learned how important it is to scope ambitious ideas into a prototype that demonstrates the core vision.

What's next for Guidely - Your indoor navigation partner

Our current prototype demonstrates the concept, but we see much bigger possibilities.

Future versions could include:

📍 Real-time AR guidance instead of static photos 🏢 Indoor navigation for schools, hospitals, malls, and airports ♿ Wheelchair-accessible routing 🌎 Multilingual voice guidance 👓 Smart glasses integration for hands-free navigation 🧠 Personalized guidance for users with cognitive disabilities 🚶 AI that understands landmarks and gives directions based on what users actually see

Ultimately, our vision is for Guidely to become an AI companion that helps people confidently navigate unfamiliar environments by understanding the real world, not just maps.