Inspiration
Travel planning starts in the wrong place. Most apps ask "Where?" but travelers ask "How do I want to feel?" We built DreamTrip to answer the real question.
What it does
DreamTrip takes emotional prompts like "I feel burnt out and need peace" and delivers ranked city recommendations, multiple parallel trip realities, and voice narration powered by ElevenLabs.
How we built it
- Frontend: React with Material UI
- Backend: FastAPI (Python)
- AI Reasoning: Google Gemini
- Voice: ElevenLabs Text to Speech, and Speech to Text
- Data: Custom city feature schema with travel archetypes
Challenges we ran into
- Handling vague prompts without breaking the pipeline
- Building rich city data from scratch
- Integrating voice without slowing the demo experience
- Creating a professional UI under hackathon time pressure
Accomplishments that we're proud of
- A working emotion to destination pipeline
- Clean professional UI that feels like a real product
- Voice narration that makes discovery magical
- Modular code that can evolve beyond hackathon
What we learned
- Ambiguity is a feature, not a bug
- Structured JSON output saves frontend integration time
- Voice makes the experience significantly more immersive
- Free text prompts work better than rigid filters
What's next for DreamTrip
- Real time pricing from Skyscanner API
- User accounts to save and revisit trips
- Mobile app with React Native
- Booking integration from dream to destination
Built With
- css
- elevenlabs-api
- fastapi
- google-gemini-api
- html
- javascript
- python
- react
- skyscannerapi

Log in or sign up for Devpost to join the conversation.