-
-
The project development enviroment, virtual studio
-
Project being tested in the device in which it was build
-
Project running locally providing user interface for testing and verifying using different prompts
-
Generated itinerary cover page
-
Generated itinerary first page
-
Generated itinerary second page
-
generated itinerary third page
-
Generated itinerary last page
🌍 TasteTrip: AI-Powered Personalized Travel Planning
💡 Inspiration
Most travel planning tools are built around menus, filters, and endless options—but real travelers don’t think that way. We think in vibes, in moods, in cravings.
I wanted to create something that listens to people. What if you could simply say:
"I love jazz, Studio Ghibli films, and French pastries"
...and the system figures out where you should go, what to do, and how to get there?
That’s what inspired me to build TasteTrip—a platform that blends personal culture with cutting-edge AI to create trips that feel truly yours.
🧭 What It Does
TasteTrip turns your cultural preferences into a day-by-day travel itinerary.
- You describe your travel taste in your own words.
- The system uses AI to extract your favorite music, food, films, and vibe.
- Qloo recommends a country, city, and a list of places based on those tastes.
- Another AI model creates a smart itinerary for your stay—morning, afternoon, and evening activities.
- If your trip is soon, you get a weather forecast.
- A photo from Unsplash gives visual inspiration.
- Directions are created with clickable QR codes for each place.
- Everything is combined into a downloadable, printable PDF.
🏗️ How I Built It
The project uses a combination of large language models and powerful APIs to turn a user's cultural preferences into a personalized, day-by-day travel plan.
Frontend: A simple input form where users enter:
- A natural-language description of their tastes (food, music, movies, vibe)
- Departure city
- Travel dates
Qloo Taste AI: The first step is to send the raw input text directly to Qloo.
- Qloo interprets the cultural context and recommends:
- A destination (country and city)
- A curated list of places to visit that match the user’s taste profile
Together AI: Once Qloo provides the destination and locations, Together AI generates a structured itinerary.
- The output includes a logical, multi-day plan with:
- Morning, afternoon, and evening activities
- A flow that reflects the user’s cultural vibe
OpenWeatherMap: If the trip is within 7 days, OpenWeatherMap is called to retrieve local forecasts for the destination.
- This helps users prepare based on weather conditions.
Unsplash: A high-resolution image representing the recommended city or country is fetched from Unsplash.
- This adds a visual touch to the itinerary.
Geoapify + QR Generator: Each place from Qloo is mapped using Geoapify.
- Coordinates are turned into QR codes that link to navigation tools
- Scannable and clickable for ease of travel
PDF Generator: All of the above data—places, itinerary, weather, images, and QR codes—are compiled into a single downloadable PDF.
- The layout is clean and optimized for offline access and sharing.
All sensitive data such as API keys are stored securely using environment variables.
🚧 Challenges I Ran Into
- Extracting clean JSON from natural language using LLMs took a lot of prompt tweaking.
- API coordination was tricky because some services depend on the output of others.
- PDF layout design needed to balance text, images, and QR codes in a readable way.
- Handling missing or vague user input required fallback logic so the app doesn't break.
🏆 Accomplishments I’m Proud Of
- Built a full travel planning experience from scratch, end to end.
- Successfully used multiple APIs together in a seamless flow.
- Created a system that feels intuitive, human, and intelligent.
- Turned vague personal tastes into a structured, beautiful, and usable output.
📚 What I Learned
- How to use large language models to convert natural language into structured data.
- How to chain multiple external APIs into a unified backend.
- How to design with the user’s journey in mind, not just the code logic.
- The importance of handling real-world ambiguity and missing data in a graceful way.
🔮 What’s Next for TasteTrip
- Add user accounts to let people save their trips and return later.
- Include flights and hotel booking options.
- Create a mobile-friendly version of the app.
- Expand the cultural input categories to include books, fashion, or social values.
- Add support for multiple languages to reach a global audience.
- Build a feedback loop so user ratings improve future recommendations.
Log in or sign up for Devpost to join the conversation.