Hearu
Hearu is a journaling app where the diary answers. Users write about their day, choose how they felt, and receive a personal reply from Bongbong, an AI character with one consistent personality and speaking style.
A normal diary stores the entry and goes quiet. A chatbot replies but leaves nothing worth rereading. Hearu keeps writing private and revisitable like a diary, and adds the one thing a diary never gives a reply.
Client
Team Side project
Year
2026
Project type
Mobile App · AI Journaling
Credits
Product Strategy · UX/UI Design · AI Experience Design
Challenge
People write in diaries and talk to AI. Neither feels complete.
More people bring their daily thoughts and worries to AI. But chat threads scroll away, and a conversation makes
a poor personal record. Diary apps have the opposite limit: they keep everything and respond to none of it.
The design challenge was to combine a diary's privacy and continuity with the immediacy of a reply, without letting Hearu feel like one more chatbot, or turn clinical and mechanical.
Research
The problem was not writing itself, but feeling heard after writing
I reviewed diary apps, AI chat services, mood trackers, and character-based companions to see
where each one loses the user. Four findings shaped the product directly.
Emotional response
Users wanted acknowledgement after writing, not only a place to record events. This became the character reply at the center of the product.
Emotional response
How might we?
Revised body copy: How might we give people the relief of being heard after writing, without giving up the privacy and continuity of a diary?
My Role
I defined the product from concept to working MVP
As the product designer, I defined what Hearu is: the concept, information architecture, journaling flow,
and an AI interaction model built around a single character.
That meant deciding how the emotion framework works, how entries and replies are archived, how privacy is handled,
and how the UI system carries users from onboarding and writing through reply generation, feedback, and looking back.
Key Contributions
01
Product concept and value proposition
02
Information architecture and navigation
03
Emotion selection framework
Key Solutions
Three moments that turn writing into a relationship
Hearu runs on one loop: Write → Reflect → Receive → Revisit. Instead of adding productivity features,
I designed each step to stay simple and lead naturally into the next.
Dashboard for writers
Outcome
The MVP completes the full journaling loop
01
Complete loop.
Write, reflect, receive, revisit
02
5 emotional states
03
2 speaking styles
04
1 consistent character
05
Integrated privacy
Reflection
A supportive AI product is more than a good reply screen
Hearu's hardest problems sat outside the visible UI. An AI journaling product has to manage response consistency,
unsafe or off-tone replies, and what happens when someone writes about a crisis.
And because it holds deeply personal data, privacy, deletion, and export are design surfaces, not legal footnotes.
The MVP covers part of this: one consistent character, tone control, reply feedback, generation and failure states,
and the PIN lock. Before scaling, the product would need safety responses for high-risk entries, clearer communication
of the AI's limits, data export and deletion controls, and a way to evaluate reply quality from user feedback.
Reply length, generation speed, and operating cost also have to stay in balance, or the service becomes repetitive
or too expensive to sustain. The next step is to validate whether Bongbong's replies are genuinely worth returning to,
while strengthening safety, privacy, and response quality before expanding.
Next project
EMS Copilot Korea
