Memora

Memora - Talk to who you used to be
Today's Entry
Past Entries
Back in Time
Chat With Your Past Self

Inspiration

We all grow, change, and forget how far we’ve come. Our journals, notes, and late-night thoughts capture who we were, but they just sit there, unread. We wanted to create something that lets people reconnect with their past selves —not just by reading, but by talking to them.

What it does

Memora lets you chat with your past self through your own written reflections. You simply type in your thoughts, memories, or old journal snippets. Memora analyzes your tone, emotions, and context, then responds as your past self would have, using your own words and patterns. It’s a reflection, reimagined as a conversation.

How we built it

Frontend: Next.js 15, React 19, TypeScript
Backend: Next.js API Routes, Prisma ORM
Database: PostgreSQL with pgvector (Supabase)
AI: Claude (Haiku 4.5 model)

Challenges we ran into

Merging the front and backend code bases
Familiarity with the Supabase database
Designing prompts that feel empathetic, not robotic.

Accomplishments that we're proud of

Built a working MVP where users can instantly chat with a version of their past selves.
Achieved natural, context-aware tone imitation using lightweight fine-tuning.
Designed a calm, journal-like interface to encourage honest reflection.
Received emotional feedback from testers who said: “It felt like my old self was actually talking to me.”

What we learned

Reflection can be made interactive when powered by personal language models.
AI empathy isn’t about perfect responses — it’s about tone, timing, and context.
Simplicity builds trust: users preferred minimal UI and human-like phrasing.
Memory and self-dialogue are powerful tools for mental clarity and growth