A designer transforming ideas into products & experiences that meet user requirements and leave a lasting impression.
About me ↗
01 · Improv · Vibe-coded app
IMPROV
Improv: Not for acting, but for stories.
Improv: Not for acting, but for stories.
Vibe Coding
Web App Design
Product Design
A tiny web app for big imaginations. Read, write one line, and leave the rest to strangers. Every story ends after just 10 lines, turning a few minutes of free time into something unexpectedly creative.
A tiny web app for big imaginations. Read, write one line, and leave the rest to strangers. Every story ends after just 10 lines, turning a few minutes of free time into something unexpectedly creative.
Timeline
June 2026
June 2026
Design scope
0 to 1
0 to 1
Design Tools:
The idea arrived while I was watching The Office. Michael Scott was going to improv class after work, and my brain wandered. What if improv lived inside an app, but people made collaborative stories instead of acting them out?
The idea arrived while I was watching The Office. Michael Scott was going to improv class after work, and my brain wandered. What if improv lived inside an app, but people made collaborative stories instead of acting them out?
01
A cat walked into a meeting with a tiny notebook.
A cat walked into a meeting with a tiny notebook.
02
Everyone pretended this was totally normal.
Everyone pretended this was totally normal.
03
Then the cat asked for feedback on its purr deck.
Then the cat asked for feedback on its purr deck.
One sentence each. Ten lines max. That is the whole game.
One sentence each. Ten lines max. That is the whole game.
Improv is a collaborative story app. One person starts a room, then each person adds one sentence. The story stops at 10 lines, because short stories are easier to finish, funnier to read, and surprisingly good at making people choose better words.
Improv is a collaborative story app. One person starts a room, then each person adds one sentence. The story stops at 10 lines, because short stories are easier to finish, funnier to read, and surprisingly good at making people choose better words.
Start
Open a room with a silly first line.
Pass
Invite others to add the next sentence.
Finish
At line 10, the story becomes a tiny keepsake.
I chose a simple idea on purpose. I wanted the first build to be memorable, not heavy.
I chose a simple idea on purpose. I wanted the first build to be memorable, not heavy.
Figma
Figma
Sketched the first screens and gave the idea a friendly shape before touching the build.
Claude Code
Claude Code
Used prompting to design, rebuild, fix, and nudge the app toward something fun.
Supabase
Supabase
Stories, rooms, and users needed a place to live. Supabase handled the backend and database.
Vercel
Vercel
Then I shipped it, because a tiny app feels better when it is actually out in the world.
Prompt note
Build a simple web app called Improv where strangers create stories together, one sentence at a time. Each person can only contribute one line, and every story ends after 10 lines. The app should feel playful, minimal, and memorable with a cute cat mascot and simple emoji reactions.
Build a simple web app called Improv where strangers create stories together, one sentence at a time. Each person can only contribute one line, and every story ends after 10 lines. The app should feel playful, minimal, and memorable with a cute cat mascot and simple emoji reactions.
The first version was rough. The cat had bigger dreams than the tools could handle.
The first version was rough. The cat had bigger dreams than the tools could handle.
It took many rounds of prompting, testing, deleting, and asking again. The useful part was learning how specific I needed to be. Vague prompts made vague screens.
It took many rounds of prompting, testing, deleting, and asking again. The useful part was learning how specific I needed to be. Vague prompts made vague screens.
I wanted the cat to feel alive.
In my head, it walked, jumped, reacted to users, and behaved like a tiny character. Current AI tools could not reliably make that happen, so I kept the mascot simpler for now.
alive, but not quite yet
The first version was rough. The cat had bigger dreams than the tools could handle.
✓ What worked
A tiny concept made the build focused. The 10 line limit gave the app a real personality. Shipping early made it feel real.
A tiny concept made the build focused. The 10 line limit gave the app a real personality. Shipping early made it feel real.
✗ What did not
The first generated UI needed a lot of care. The mascot idea also needed more control than the tools gave me.
The first generated UI needed a lot of care. The mascot idea also needed more control than the tools gave me.
☘︎ What I learned
Vibe coding feels fastest when the idea has edges. A small app can teach a lot when you design, build, and ship it yourself.
Vibe coding feels fastest when the idea has edges. A small app can teach a lot when you design, build, and ship it yourself.
I wanted the cat to feel alive.
In my head, it walked, jumped, reacted to users, and behaved like a tiny character. Current AI tools could not reliably make that happen, so I kept the mascot simpler for now.
alive, but not quite yet
I wanted the cat to feel alive.
In my head, it walked, jumped, reacted to users, and behaved like a tiny character. Current AI tools could not reliably make that happen, so I kept the mascot simpler for now.
alive, but not quite yet
Improv is currently being shared across social media to grow its user base. As people begin using the app, this page will be updated with engagement and usage metrics.
Improv is currently being shared across social media to grow its user base. As people begin using the app, this page will be updated with engagement and usage metrics.
←
←
Previous
Previous
Next
Next
