Eppoi — a bedtime story from a parent's day

Three app screens on a dark blue field, connected by a hand-drawn yellow line that turns into a rocket: your day is full of events, we help you build stories, to tell at night to your child.

The problem

Every night, a parent has to invent a story. Tiredness isn't the hard part. Inventing on demand is a creative task, and it comes due every single day.

The existing alternatives solve the time problem but remove the relationship. A book tells the story for you. A video replaces you entirely.

How do you give a parent the structure of a story without taking away their voice?

An ordinary day is already a story. It's only missing the grammar.

Alarm, 7:00: The sun rose over the kingdom.
Leaving home, 8:00: Milla the pirate set sail on the Galleon.
Lunch, 1:00: The spaceship stopped to refuel.
Email, 6:00: The pigeon crossed the ocean.
Traffic, 6:30: A meteor storm slowed everything down.

Research

56 responses to an online survey, plus a round of interviews that produced five personas.

52%

already invent stories rather than read them

56%

say they struggle to tell a story

41%

say their own day strongly shapes what they tell

68% at least moderately

n = 56 · online survey and interviews

Half of these parents already invent rather than read. The intent is there, and it isn't a question of willingness.

But more than half say they get stuck, and the open answers name the same failure in different words: losing the thread when interrupted, stopping and starting over, not managing to hold the story together to the end.

“Sometimes I lack imagination; I pause to think, and then I start again.”
Marco C.

“The difficulty lies in remaining consistent while I tell the story.”
Francesca P.

“When I get interrupted, I lose the thread of the conversation.”
Giovanni P.

“maintain a narrative logic”
Caterina M.

The problem wasn't imagination. It was grammar.

A map of existing alternatives on two axes, static to interactive and standard to personalized. Story books, flashcards and video platforms sit on the static side; voice assistants are interactive but standard. The interactive and personalized quadrant is empty except for Eppoi.

Books tell the story for you. Videos replace you. Voice assistants personalize but take the voice away. The interactive-and-personalized quadrant was empty, and filling it meant the parent had to stay the author.

Serena

“Nothing is perfect, everything can be improved.”

31 · Pastry chef · Rome

She became a mother at a difficult, unplanned moment. She doesn't trust her own instincts, so she raises her son by strict rules she isn't happy with. She knows something has to change.

Wants: To change how she parents
Goal: A calmer, more relaxed child
Friction: No tools to do it with

The core user. The product exists for her.

Paola and Giuseppe

“One life, one long journey.”

72 and 75 · Retired · Genoa

They visit their grandchildren, who live in another city. They love collecting the stories of people they meet and telling their own — starting with how the two of them met, on a trip back at university.

Wants: To pass on what they've lived
Goal: To hold a child's attention
Friction: Distance, and using a device at all

The pair that stress-tests the product: a screen is the obstacle, not the answer.

The storyteller isn't always a parent: grandparents far away, fathers who see their children on weekends.

The system

Propp's cards already existed as an analog game: 31 narrative functions that Gianni Rodari had turned into a tool for inventing stories. I didn't invent the mechanism. I gave it a context.

The phases

Propp's 31 functions, grouped

The 31 functions grouped into four phases: initial balance, 10 functions; breaking the balance, 6; the hero's trials, 11; resolution, 4, not used in the app.

The help is closed by default. A parent who knows what to say never opens it.

One card per function

Three layers, from the drawing to the help

A single card: the title of the function, an illustration, the sentence that describes it, and a collapsed help bar at the bottom.

The help is closed by default. A parent who knows what to say never opens it.

What's left

The output isn't a text

The output: a waveform labelled the parent's voice. The recorded story stays in the collection and can be played again.

The product never tells the story for you.

Two ways in

The same 31 functions, reached from two different starting points. Only the fourth step changes.

Mode 1

Daily experience

Source

Vladimir Propp

Linguist. Mapped the recurring structure of the folk tale.
Phases

4 phases

Initial balance · Breaking · Trials · Resolution
Functions

31 functions

Departure, prohibition, the magical gift, the struggle…
The difference

Everyday situations

Waking up, the commute, the groceries, getting home late.
Output

The app

You start from what actually happened today. The day is the way in, and the functions arrive already dressed as your own life.

Mode 2

Propp's laboratory

Source

Vladimir Propp

Linguist. Mapped the recurring structure of the folk tale.
Phases

4 phases

Initial balance · Breaking · Trials · Resolution
Functions

31 functions

Departure, prohibition, the magical gift, the struggle…
The difference
Straight back into phases

No anchor to the day.
- Initial balance
- Breaking
- Trials
Output

The app

Pure story, with no day behind it. For the nights when nothing happened worth telling — and for whoever wasn't there that day.

Four decisions

"Using a device during an intimate moment" — the weakness I named in my own SWOT, and the constraint that shaped everything after it.

I started from Propp's cards, an analog tool that already existed.

Alternative: Inventing a story-generation mechanism from scratch.
Why: What the research showed missing was structure, not imagination. And the structure already existed — just not in the hands of a parent at nine in the evening.

The help sits under the card, not on top of it.

Alternative: Showing prompts and ready-made phrases up front.
Why: An app that fills the screen with text makes a parent read instead of look at their child. The help opens only when it's needed, and anyone who knows what to say never opens it.

Every card carries questions to ask the child, not just lines to say.

Alternative: Giving the parent narrative prompts only.
Why: The real need wasn't telling better stories, it was time spent together. Questions move the child from listener to co-author.

Two modes: the day, or the laboratory.

Alternative: A single mode, anchored to the day just lived.
Why: There isn't something worth telling every night, and not everyone who tells the story spent the day with the child. Same narrative engine, two ways in.

The flow

From the day just lived to a recorded voice, in five steps.

What I didn't validate

I used the system myself to build complete stories, checking that the cards produced coherent narratives from initial balance through to resolution. The mechanism works on paper.

What I didn't validate is the part that matters most: whether it works at nine at night, with a tired child and a parent who has already given everything. It's the same tension I named in the SWOT. The collapsible help panel is my design answer to that risk, but it stays an untested hypothesis.

Visual system

One hand-drawn line, four colors, and a language a child recognizes before they can read.

Eppoi turns a parent's day into a bedtime story.