← Back to work
Consumer App 2026 · Solo Designer-Builder · In Beta

A quiet practice of kindness.

A daily kindness app: three small acts a day. No streaks. No badges. No guilt. Framed, designed, and shipped solo, using modern tooling without losing rigor.

Role
Solo designer-builder
design · content · ship
Timeline
~2 weeks
active build · evenings & weekends
Stack
Figma · Lovable
Claude Code · Supabase · Git · Vercel
Status
Beta · 50 testers
mixed contexts
50Beta testers
150+Acts & prompts written
100%Solo · no engineer
~1 minTo first act
·In motion

The product, moving.

The morning offering, start to finish, screen-recorded from the live beta.

Flow 1 the morning offering, start to finish
01The problem

Every habit app runs on guilt. Kindness can't.

A kindness practice has to feel like a gift to return to, not a duty to perform. The entire habit-app genre runs on a mechanic this category simply can't survive.

The category

Streaks. Badges. Shame.

Habit apps work by making you feel a little bit behind. A streak you can break. A badge you haven't earned. A red dot demanding return.

The contradiction

You can't guilt someone into being kind.

The moment kindness becomes an obligation, it stops being kindness. The whole genre runs on a mechanic this category can't survive.

The thesis

A kindness practice has to feel like a gift to return to,
not a duty to perform.

02Five decisions

Holding the thesis across every screen.

Five moments where the easier path lost, and what the thesis cost in return.

01The foundational choice
No streaks. No counts. No scoreboard.

The decision that forced every other decision.

Considered

A soft streak. A weekly rhythm. A quiet personal count of completed acts.

Chose

None of it. No numbers anywhere. The garden grows visually: users see what they've done, never how much.

Tradeoff

Retention may suffer. Protecting the meaning of the practice is the thesis. That's the right cost.

02Systems thinking
A five-tier system that respects the user's state.

Not a difficulty curve. Five different kinds of kindness: each one right for a different kind of day.

Microalmost nothing
Gentlea soft reach
Bolda real reach
Deepan inward turn
Bravecourage

Library mix: 20% / 35% / 25% / 12% / 8%. Most days, most people want gentle or bold. Brave stays rare on purpose.

03Architectural rigor
A matching cascade, not an algorithm.

When AI would have been easy and wrong, I chose rules. Every result traceable.

i.tier + audience match
ii.tier only
iii.audience only
iv. ✓universal: the safety net

What it saved$50–200/mo in API costs · full control of voice · zero unpredictable outputs in a sensitive product.

04Knowing what to cut
A marketing site I almost shipped, and cut.

Wrong tool, right product, wrong moment.

Designed

A six-section marketing site, 1,800 words, ready to deploy.

Shipped instead

8 in-app discovery cards, opened from a small prompt below the day's offerings.

Why

Marketing sites acquire users. I wasn't acquiring. I was in invite-only beta. Users needed the philosophy where they actually were.

05Restraint as craft
Five rounds on one card colour.

Restraint scales. Cleverness doesn't.

i.Uniform warmth no rhythm
ii.Whisper tint per tier too subtle
iii.Soft tier tints starts to feel busy
iv.Multi-direction gradients read as a bug
v.Single warm gradient, identical on every card shipped

The clarity of thought, and the journey you took me on in those five minutes was very, very special. I loved it as a designer, and more as a user.

A designer · beta tester
Flow 2 completing an act on the phone
·The garden

What grows instead of a number.

Every act plants something. The garden is the only record: you see what you've done, never how much. Testers said it became the part they show friends.

The Sprig garden: a real visual record of completed acts
The garden: a visual record, no counts, no scores
03The visual language

A system designed to disappear.

Warm, restrained, generous with space. Distinctive type does the emotional work; everything else steps back.

Palette
Cream, coral, mostly cream.
bg
paper
ink
accent
leaf
Typography
Serif italic for the soft moments.
a quiet practice
Instrument Serif · titles, emotional moments
text someone you miss
Fraunces · act titles, body italics
today's offering
DM Sans · UI labels, metadata
04Three voices, one screen

The act-detail page is a small system.

Three content tables collaborate to make one quiet moment: lead-in, act, reflection.

Before reflective_prompts

“this one asks for a little courage. you have it.”

A one-line lead-in chosen by the cascade. Settles the user before the act.

The act acts

“tell a friend specifically what you admire about them.”

Lowercase, imperative, concrete. The thing the user actually does.

After sprig_responses

“what did it feel like, in the moment you actually did it?”

Reflection: an open question or warm affirmation. Closes the loop.

Try it · live in beta

Open Sprig on your phone.

It's live, it's free, and it works in any browser. The morning flow takes about a minute. Three small acts will be waiting.

Web app No signup wall ~1 min to first act
Open Sprig ↗
sprigkindnessapp.vercel.app
QR code to open Sprig
scan to open on your phone
05The build

Solo, with AI as execution partner.

Tools handled execution. Every design decision was mine.

designFigma
UI scaffoldLovable
surgical editsClaude Code
data + authSupabase
hostingVercel
~2 weeks
design to beta
150+
acts, prompts & responses written
100%
solo · no engineer
06What I'm learning

Honest signal from real users.

No invented metrics. Three weeks in beta: here's what testers are telling me.

What landed
The voice is the product.

Testers consistently mention the tone. Several said variations of “it doesn't feel like an app.” The thesis returning as a quote is the best signal a designer can get.

What landed
The garden became social.

I designed the garden as a quiet personal record. Several testers said it's the part they show friends, earning its keep visually, even without metrics.

What didn't
Onboarding was too eager.

The about-cards are dense, and several users skipped them. v2 will start with three cards, with deeper philosophy unlocking progressively as users return.

What didn't
“No streaks” confused.

A handful of testers asked “how do I track my progress?” The absence isn't legible to people raised on streak-driven apps. Experimenting with a “your week so far” record. No count.

What's next
Three open questions for v2.
Does no-streaks hurt retention, and if so, can return be supported without re-introducing guilt?
Are users avoiding brave acts because brave is hard, or because they surface at the wrong moment? A matching problem, or a courage problem?
What's the right release rhythm for new content: a slow drip, or an open library?
What it proves

I can do the work, alone.

I can frame a problem nobody handed me. Hold a thesis across hundreds of decisions. Design a system, not just screens. Ship something real, solo, using modern tooling without losing rigor.

What it doesn't

How I collaborate.

No PM, no engineer pushing back, no design director critiquing. I know I can do the work alone. I'm looking for a team where I can do it alongside others.

Let's work together

Actively looking for senior
design roles in the Bay Area.