Epiphany

Sign in
Bullet journal:A drag-and-drop bullet journal lets you scrapbook your notes into a spread with stickers, handwritten text, and photos
To-dos: Any note that sounds like a task gets automatically pulled out into a dedicated to-do list
Home / stream: Notes are auto-sorted into folders and dropped into a stream with a generated instrumental playing based on your recent mood

Inspiration

Great ideas rarely arrive when we're sitting down to work. They happen while walking to class, lying in bed, listening to music, or in the midst of a stakeholder meeting. The problem is that these thoughts are often forgotten or end up scattered across notes apps, voice memos, and sticky notes.

We built Epiphany to solve this problem. Instead of simply storing notes, Epiphany helps users capture fleeting thoughts instantly and transforms them into personalized reminders, recommendations, and creative experiences. By combining quick note-taking with AI-powered organization, Epiphany transforms random ideas into actionable plans, personalized suggestions, and tailor-made jingles.

Inspired by the creativity of scrapbooks and bullet journals, Epiphany gives users a space to decorate, organize, and revisit their thoughts in a meaningful way. We wanted to move past the constraints of a productivity tool. We wanted to build a place where ideas could become memories, plans, and sources of inspiration.

What it does

Epiphany is an AI-powered scrapbook that makes capturing ideas effortless.

Users can quickly record thoughts through text or voice, without worrying about organization. Using Gemini AI, Epiphany automatically:

Categorizes thoughts into meaningful topics.
Generates personalized to-do lists based on your notes.
Suggests recommendations and next steps tailored to your interests and goals.
Keeps your thoughts organized so nothing important gets lost.
Encourages creativity by allowing users to bullet-journal their thoughts and customize their digital scrapbook.
Uses AI to transform thoughts into personalized jingles, giving users a unique way to experience and remember their ideas.

Rather than forcing users to manually sort notes into folders, Epiphany understands the context of each thought and organizes everything intelligently. Gemini acts as the "brain" of Epiphany by analyzing every captured thought, extracting meaningful information, identifying actionable items, and generating personalized recommendations instead of simply storing text.

How we built it

We built Epiphany using a modern full-stack web architecture.

Frontend: Next.js, React, and TypeScript
Backend: Next.js API Routes
Database: MongoDB
Authentication: Secure user authentication for personalized accounts
AI: Google Gemini API for intelligent categorization, recommendation generation, and personalized to-do lists
Voice & Audio: ElevenLabs for voice-to-text capabilities and transforming thoughts into personalized music

Google Gemini Gemini serves as the "brain" of Epiphany by analyzing every captured thought, extracting meaningful information, and providing useful suggestions instead of just storing text. Through prompt engineering, we designed Gemini workflows that can understand the user's intentions, identify important tasks, and generate personalized responses based on individual thoughts.

ElevenLabs ElevenLabs handles speech-to-text for every voice recording before it's classified, and powers generative music: each note's mood is mapped to a tempo and prompt, then used to generate a short original instrumental via the ElevenLabs Music API.

MongoDB Atlas Three collections (users, notes, and categories) store all app data, with indexes designed around the app's actual query patterns for fast per-user lookups and auth. Categories aren't pre-seeded, but created dynamically as Gemini generates them, with a unique index preventing duplicates.

Challenges we ran into

One of our biggest challenges was designing prompts that consistently produced useful and structured AI responses. We wanted recommendations that felt personalized rather than generic, while ensuring Gemini could accurately understand the context behind each user's thoughts.

Another challenge was integrating the AI workflow into a responsive user experience while keeping note creation quick and seamless. Balancing speed, usability, creativity, and intelligent processing required several iterations.

We also spent time designing a database structure that could efficiently store user thoughts, AI-generated metadata, recommendations, and creative content generated from AI services.

Accomplishments that we're proud of

Successfully integrating Gemini AI to automatically categorize thoughts and understand user intent.
Creating personalized to-do lists and recommendations directly from users' ideas.
Designing a scrapbook-inspired interface that encourages users to creatively express themselves.
Developing bullet-journal features that make organizing thoughts interactive and enjoyable.
Developing a full-stack application with secure authentication and persistent storage through MongoDB.
Using ElevenLabs to transform thoughts into personalized music and create a more immersive journaling experience.

What we learned

Throughout this project, we learned how to build AI-first applications rather than simply adding AI as an extra feature.

We gained experience working with prompt engineering, integrating large language models into production workflows, managing structured AI outputs, and connecting frontend interactions with backend AI processing. We also improved our skills with Next.js, MongoDB, and full-stack application development.

Most importantly, we learned how AI can be used as a creative collaborator — not just a productivity assistant. By combining AI organization, recommendations, and generative audio with a creative journaling experience, we were able to build a tool that helps users capture and transform their ideas.