SkyVisionAI

Visualize your next destination before you even take off. Our AI transforms your preferences into visual art to recommend the three countries that best suit you. Less doubt, more world.

Comment

Inspiration

Our inspiration is simple: we love to travel, but we hate analysis paralysis. Spending hours jumping from tab to tab trying to find a country that matches what we feel is exhausting. That's why we decided to reverse the process. Instead of choosing a name on a map, we invite you to describe what you're longing for. We use technology to shape your ideas and give you back three real destinations. Because if you can imagine it, there's a place in the world waiting for you.

What it does

Our tool acts as a dream translator. You enter the keywords for your ideal trip, and we generate a personalized image using AI. But it doesn't stop at art: we analyze that image to connect you with three real destinations around the world, giving you the final push you need to choose your next trip. And we don't just show you where to go. Aura Travel connects with real-time flight data so you can see transport options and costs instantly, turning your visual inspiration into an actionable travel plan.

How we built it

AI Engine: We used Google AI Studio to interpret user desires, generating both the destination logic and the visual descriptions. Travel Integration: We integrated the Skyscanner API to transform visual inspiration into real flight options and costs. Technical Pivot: While we initially developed on Localhost, we encountered CORS errors with the Skyscanner API and hit image generation limits. To overcome this, we built a refined prototype that showcases the high-quality results from Google AI Studio, ensuring a smooth and functional demonstration of the concept. Built with JavaScript, focusing on a clean interface that connects the user's "vibe" with tangible data.

Challenges we ran into

Prompt Engineering: Adjusting the AI ​​so that the user's (sometimes contradictory) keywords would generate a coherent image and not visual chaos.

Translating the Visual into the Real: The biggest challenge was creating an algorithm that could "understand" the generated image and accurately associate it with real countries. It's not easy for code to understand that a snow-covered mountain with temples points to Japan or Switzerland.

API Synchronization: Ensuring that image generation and destination search occur within an acceptable waiting time so that the user does not lose interest.

Accomplishments that we're proud of

The "Magic" of the Result: Making the generated image truly capture the essence of what the user had in mind. Seeing abstract concepts transform into a tangible landscape is incredible.

What we learned

Data Curation: We discovered that to recommend destinations, geographical labels are not enough; we need to understand the "vibe" of each country (climate, architecture, atmosphere).

What's next for SkyVisionAI

The next for Aura Travel is to implement a map, so that while the user enters words and creates the image, there is simultaneously a world map that moves the user according to the words they enter.

Built With

Share this project:

Updates