We all have dreams that are absolutely unhinged, oddly profound, or just laugh-out-loud funny—but they slip away from memory the moment we open our eyes. What if we could capture that fleeting chaos and turn it into something worth sharing?
Dreamify was born from the idea that dreams are mini-stories worth telling. They're bizarre, hilarious, sometimes emotional—and uniquely you. Sharing them with friends lets others in on your subconscious adventures, sparking conversations, laughter, and connection, no matter the distance or time. It’s like building a diary, but one that your friends can read, enjoy, and react to.
We wanted to create a space where people can stay connected through the weirdest and most personal parts of life in the most fun way possible.
Dreamify is your personal dream-to-story machine with a social twist:
- Users start by jotting a quick dream recap when they wake up.
- We feed it through a custom AI pipeline using OpenAI's LLM, engineered with carefully crafted prompts to produce short, funny, and engaging summaries.
- Then, based on the story, we generate dream-inspired images using a Stable Diffusion model.
- These dream stories and visuals are then posted to a user’s page.
- Friends can like, comment, and view each other’s dreams, building a social platform driven by the subconscious mind.
It’s your dream journal, but funnier—and social.
This was truly a 24-hour full-stack learning journey:
- Frontend: Built in React Native with Expo—none of us had used React Native before this hackathon!
- AI:
- We used OpenAI’s GPT models for text generation via prompt engineering.
- Story titles and image prompts are automatically created.
- Stable Diffusion (local model) was used to generate relevant, creative, and sometimes hilarious images based on the dream stories.
- Backend:
- Developed in Python using FastAPI
- SQLite database and REST APIs
- Tools: Postman, Uvicorn for local dev server
We engineered a full content pipeline from dream to story to shareable post.
- Getting all the AI dependencies (especially for Stable Diffusion) installed correctly - we needed to use GPUs!
- FastAPI endpoints took trial and error to get running smoothly.
- Typescript issues popped up constantly… and yes, centering that one tiny navigation highlight might’ve been our longest task 🫠.
- React Native was a whole new world and getting comfortable with layouts, styling, and state management took time.
- A functional full-stack mobile app built in 24 hours!
- Clean, aesthetic UI with page navigation and functional social features.
- Seamless AI integration for text and image generation.
- Most importantly: turning our wild concept into something real, creative, and shareable.
- Everything. No really—everything.
- None of us had built a mobile app before.
- Learned how to:
- Use React Native and Expo from scratch
- Write, connect, and test APIs
- Handle backend and frontend integration
- Work with and optimize AI models for creative use
- Streamlining the story/image generation pipeline to reduce lag
- Adding user auth and multi-user testing
- More social features like replies, following, and tagging
- Speech-to-text generation to enter dream recaps.
- Possibly: dream streaks, personality-based dream analysis, or dream maps over time!
Dreamify has the potential to become a meaningful and delightful daily ritual for users—and we’re excited to keep dreaming it into reality 🌙