Ikigai

Inspiration

We live in an era of too much noise. Job rejections, AI anxiety, phone addiction, losing patience — people are overwhelmed and don't know which direction to go. Most wellness apps respond by measuring you: steps, calories, sleep scores. I wanted to build something different. Ikigai — the Japanese concept of "reason for being" — became the answer. Not an app that judges you. An app that reflects you.

What it does

Ikigai is a mood-aware AI wisdom experience with three personas that automatically respond based on your emotional state. Feeling lost or heavy? The Elder appears — warm, unhurried, like a grandfather sharing tea on an autumn afternoon. Excited to build something? The Kid shows up — fearless, joyful, ready to start. Searching for answers? The Wanderer walks alongside you, asking the right questions. The UI shifts with your mood — seasonal skies and painterly backgrounds change automatically. Autumn for heavy days. Spring for hopeful ones. Summer for searching ones.

How we built it

Frontend: Next.js deployed on Vercel, scaffolded with v0 AI: Groq (llama-3.3-70b-versatile) with a custom mood routing agent that classifies emotional intent before selecting a persona Database: AWS DynamoDB for real-time conversation storage, AWS S3 for archiving full journey files as JSON Auth: AWS IAM with least-privilege access keys Each message triggers two API calls — one to classify mood, one to generate a persona response — then saves to both DynamoDB and S3

Challenges we ran into

Getting the mood routing agent to reliably classify emotional intent across varied phrasing Wiring the frontend persona switching so the UI updates automatically based on the API response Managing the dual-storage pattern across DynamoDB and S3 without slowing the response time

Accomplishments that we're proud of

The UI genuinely feels like a living world — not a chatbot interface The mood detection works — typing "I want to build something" actually switches to the Kid, with the sky changing to spring DynamoDB and S3 both saving every conversation in production Going from a sketch on paper to a fully deployed app in under two weeks

What we learned

How to build a mood-aware multi-agent routing system from scratch without LangChain AWS dual-storage architecture: DynamoDB for fast queries, S3 for archival and future analytics How to use v0 for rapid UI scaffolding while maintaining full control of the backend That the best product ideas come from personal pain — I built this because I needed it

What's next for Ikigai

Mood Journey Dashboard: a /journey page showing your emotional patterns over time — which personas you visited, your current "season," whether you're trending toward spring or stuck in winter Crisis Detection: gentle surfacing of crisis resources when distress signals are detected Voice Mode: speak to the Elder instead of typing Memory Across Sessions: the Elder remembers what you shared last time Ikigai Score: after enough conversations, Ikigai reflects your emotional arc back as a season, not a number

Built With

• amazon-web-services
• aws-dynamodb
• aws-iam
• css
• groq-ai
• next.js
• tailwind
• typescript
• v0
• vercel