Curiosity Cloud: a learning companion built for six-year-old thumbs, not adult ones.
A single, flat, icon-first home for a bilingual primary school network — where activities, stories, art, and sound live together, and no child needs to read a word to find their way in.
One companion, two very different readers.
The brief came from a bilingual school network that wanted a single digital space where children from Grade 1 through Grade 6 could reach their classroom activities, storybooks, recordings, posters, and their own drawings — without a parent or teacher opening it for them. The hard part wasn't the content. It was designing one home screen that worked for a six-year-old who can't read yet, and a ten-year-old who's already bored of anything that looks "babyish."
What I owned
End-to-end UX: field research in two cities, persona synthesis, the information architecture, low- and high-fidelity wireframes, and the visual direction handed to the illustration and development teams.
What made it hard
Two age bands, two literacy levels, two languages, and a non-negotiable constraint: nothing that resembles an ad, a score, or a leaderboard — this had to feel like a classroom, not an app store.
Research, concept, design, and a lot of paper before any screen.
The same six-stage backbone I use on every product, adapted here for an audience that can't fill out a survey or think aloud the way an adult can.
Discovery
School visits, teacher interviews, classroom observation in Madrid and Barcelona.
Concept
Persona synthesis, flat IA sketching, early "point and show me" paper tests.
Design
Wireframes, icon-first navigation, visual direction for illustration.
Prototype
Clickable paper and low-fi digital mock-ups tested in real classrooms.
Validate
Formative testing with children, teachers, and parents at every round.
Four people this had to work for.
Field visits and teacher conversations across two schools surfaced four very different relationships to the same screen. None of these are the real children or teachers we met — they're composites built from patterns across many conversations.
"I want to finish my work fast so I can draw."
- Wakes early, loves quiet mornings and drawing before school
- Wants her art noticed, not just graded
- Curious about English through a favourite aunt abroad
"I only know it's mine if it has a picture, not words."
- New to the school; anxious about being apart from his brother
- Can't yet read menu labels — needs icons and sound
- Responds strongly to gentle, specific encouragement
"I need to see who's stuck without opening ten reports."
- Teaches English, art, and craft to mixed-ability classes
- Wants to introduce new techniques without a steep learning curve
- Fascinated by how differently each child approaches the same task
"I want to know she's okay, not watch her every click."
- Wants visibility into progress without feeling like surveillance
- Not part of the original brief — added after classroom visits
- Represents a stakeholder every kids' product quietly has
Flatter than the brief asked for.
The original request listed nine separate destinations. For a reader who can't yet read, nine unlabelled choices is nine chances to get lost. I merged categories that shared an instinct — art and journalling both mean "make something," video and audio both mean "sit back and watch or listen" — and added one destination nobody asked for.
(activities)
(art + journal)
(video + audio)
(posters + docs)
(recommendations)
(new — for Elena)
Nine destinations became six. "Family Hub" didn't exist in the original brief — it came directly out of a conversation with a parent during classroom visits.
Motivation without competition.
The brief didn't ask for gamification, but a learning product for children rarely survives without some sense of momentum. The layer I proposed avoids scores and leaderboards entirely — the goal is a calm sense of "I did something," not "I beat someone."
The Star Map
Finished activities light up a personal constellation instead of a progress bar. Nothing is ever marked wrong — a star is either lit or "not yet."
A Companion, Not a Score
A small on-screen creature reacts warmly to what a child does. No points, no ranking against classmates.
Voice-First Navigation
Every icon and label can be tapped to hear itself read aloud — essential for a Grade 1 reader like Mateo.
Invisible Difficulty
Activities quietly adjust after repeated attempts. There is no visible "level," so no child feels behind in public.
The Class Wall
A teacher-moderated space where Studio entries can be pinned — an audience for Amara's art, without opening it to the internet.
Calm by Design
No ads, no external links, muted transitions instead of flashing rewards — built to hold attention gently, not compulsively.
Reimagined, not redrawn.
These are fresh mock-ups built for this write-up — new colours, new icons, new layout logic — expressing the same underlying decisions the original wireframes made.
Sequence Path — logic, without a single word to read.
Looking back at this project with fresh eyes, the picture-word match alone doesn't stretch Amara and her Grade 4 classmates much. A look across today's most recommended learning apps — Wirecutter's roundup of tablet learning apps among them — points at a consistent pattern: the apps that hold up best either let kids build with icon-only logic blocks (as ScratchJr does for early readers) or stay deliberately open-ended, with no score or clock pressuring the child. Swift Playgrounds, by contrast, needs a third-grade reading level to use alone — exactly the barrier Curiosity Cloud was built to avoid. Sequence Path borrows the block idea and strips it down further: no words on any block, just icons, made for the Explore tab's older-grade track.
No text, ever
Every block is a pictogram — forward, turn, repeat — so a non-reader can build the same sequence a Grade 4 child can.
Open-ended, not timed
Matching the "sandbox app" pattern favoured by researchers: no clock, no fail state — just try again until the fox reaches the star.
Grows with the child
Grade 1 gets two blocks and a short path. By Grade 4, the same activity adds loops and longer paths — one interface, two reading levels.
You can't run a survey on a six-year-old.
Every method here was adapted for an audience that can't self-report the way an adult participant can.
- "Point and show me" sessions — instead of asking children to describe a screen, we asked them to point at where they'd go, and watched.
- Paper prototypes tested in real classrooms before a single screen was built, cutting expensive rework later.
- Teacher shadowing across full class periods to see where the paper flow broke down in practice.
- Parent surveys for at-home usage patterns — capturing what happens outside the classroom, unsupervised.
- A colour-blindness and legibility pass on every icon and tile, given the age group's limited ability to describe visual confusion.
What this project taught me.
Designing for children reset almost everything I assumed about "intuitive" design. An interface that feels obvious to an adult can be completely opaque to a six-year-old who can't yet read a label — and a nine-year-old will reject anything that feels like it was designed for someone younger. The real work wasn't the grid of icons; it was earning the trust of two age bands, two languages, and a teacher who had to make it work in a room of twenty-five different attention spans at once. If I carried one decision forward into every project since, it's this: cut the destinations until a child could get lost on purpose and still find their way home.