Case study · 04

MimoSE

Look back, make sense of me — an AI companion for self-reflection.

Visit sitemimose.io.vn

Role: Product · Design · Full-Stack
Timeline: Aug 2025 – Present
Stack: React · Spring Boot · OpenRouter
Status: Live Product

mimose.io.vn

01 · Context & Users

People who want to understand themselves better through daily self-reflection — in English or Vietnamese.

02 · The Problem

“Logging emotions every day is hard — it is easy to give up without something bringing you back to it.”

03 · My Role

Independent build: I own the product end to end — direction, UX, the React frontend and the Spring Boot API.

Product
UX Design
React Frontend
Spring Boot Backend
AI Prompt Design

04 · UX & Product Decisions

Inspired by Reflectly, MimoSE went through four phases: a garden-metaphor journal, energy tracking, a pivot to an AI companion with a relationship map — and most recently a single “Today” screen instead of four tabs. Along the way it was briefly called Aura Self AI, then came back to its first name, MimoSE — Make Sense Of ME — as a station of my Time Machine, with a Mimo of its own: not the one who drives the machine, but one who listens while you talk, then asks the next question.

Decision 01
Reminders first
~~Full journal on day one~~ — build the habit before features
Decision 02
Mimo, a companion
~~A coach persona~~ — reflection, not instruction
Decision 03
One “Today” screen
~~Four tabs + a chat button~~ — what needs attention today, in one place

05 · Technical Solution

A React 19 single-page app and a Spring Boot 3.5 API on PostgreSQL. Mimo’s chat runs through OpenRouter; each conversation is summarised into insights you can keep in your Mirror (a Johari Window), a note, or a person on your relationship map.

Client
React 19 · Vite
API
Spring Boot 3.5
AI
OpenRouter
Tools
Mirror · People

06 · Measurable Results

The numbers

Source: project repositories · Oct 2026

4
Product Phases
From journal to AI companion
350+
Commits
Frontend + backend, since Aug 2025
2
Languages
Full English / Vietnamese parity

Commits per quarter

Build activity across the frontend and backend repositories.

47
Q3 ’25
143
Q4 ’25
52
Q1 ’26
1
Q2 ’26
115
Q3 ’26

07 · Lessons Learned

What I learned, and what I would do differently next time.

01
Habit before features
The reminder that brought people back mattered more than any feature added later. Retention starts with returning, not with options.
02
A stable data model makes pivots cheap
The latest redesign changed navigation, names and UX across the app without changing a single API endpoint.
03
Dead features still cost attention
Energy tracking and action protocols still live in the backend after the pivot. Next time I would remove them as soon as the new direction is confirmed.

Have a similar problem to solve?

I take on remote freelance work as a Product Engineer — from clarifying the user problem to shipping the product.

Let's talk →

Next case study · 01

Harnix →

UX, SEO and messaging that sell AI to enterprise buyers.

04 · UX & Product Decisions

Decision 01

Reminders first

~~Full journal on day one~~ — build the habit before features

Decision 02

Mimo, a companion

~~A coach persona~~ — reflection, not instruction

Decision 03

One “Today” screen

~~Four tabs + a chat button~~ — what needs attention today, in one place

05 · Technical Solution

Client

React 19 · Vite

API

Spring Boot 3.5

OpenRouter

Tools

Mirror · People

07 · Lessons Learned

What I learned, and what I would do differently next time.

Habit before features

The reminder that brought people back mattered more than any feature added later. Retention starts with returning, not with options.

A stable data model makes pivots cheap

The latest redesign changed navigation, names and UX across the app without changing a single API endpoint.

Dead features still cost attention

Energy tracking and action protocols still live in the backend after the pivot. Next time I would remove them as soon as the new direction is confirmed.