Why a father of three built an AI that writes bedtime stories where your kids are the heroes
It was 7:45 on a Tuesday night when Dan turned his laptop around to show me what he called “the bedtime trap.” I had stopped by to pick up a car seat, and instead I got a demo. His middle kid, Zoey, was already in pajamas, and from down the hall I could hear her repeat the same phrase every parent knows: “Dad, are we doing stories tonight or not?” She said it like a threat, not a question.
Dan is a software engineer with three kids under seven. He is not the kind of parent who uses technology to skip the hard parts. If anything, he over-engineers them. When his youngest went through a phase of refusing to even sit for a book, Dan did not push it. He opened a laptop, wrote some code over a few weekends, and built a small tool that writes a bedtime story from a few short prompts — and puts the actual child in the lead role. Six years old, likes dragons, afraid of thunderstorms. That is all it takes. Zoey’s name shows up in the second sentence, and from then on, she is the one who finds the dragon’s lost egg, convinces the grumpy wizard to share the warm jam, and arrives home just before the rain starts.
It was the rain detail that got me. Dan had typed in that she was scared of thunder. The generated story did not make her brave and loud. It made her the one who notices the weather before anyone else, which is exactly how Zoey is. In that sentence, she was not a fake fantasy hero. She was a six-year-old who knows the sky and knows a way back home. That was the first moment I understood why he built this thing — and why he wanted parents to try it for free for two weeks, a beta with no credit card. I thought it was a clever gimmick until the demo was over. Then I wanted to understand the real reason it works.
Reading experts will tell you that a child who appears in a story pays attention differently than a child who silently listens. That is true, but I think the sharper insight sits in an awkward place. A young child spends most of the day being told what to do. The point of a story like this is not self-esteem. It is ownership. When your name is on the page, the story asks something of you: “What will you do?” Not “what will the bear do?” The child’s mind switches from receiving a lesson to rehearsing a decision. That is a different mental muscle. It is less about reading and more about thinking, yet it happens while they are reading.
I tried it with my own daughter the next night. She is five, and her combination of shyness and stubbornness has always made bedtime unpredictable. I typed three things into the box: her name, that she has a stuffed elephant named Shorty, and that lately she gets mad when it is time to leave the park. That last one was not a safe, sweet detail. It was real.
The story the tool wrote was not a lecture. She was an explorer who found a park deep in the woods, with a gate that would not unlock unless she said goodbye to each tree, one at a time. My daughter listened with the kind of quiet that makes a parent suspicious. Then she sat up and asked: “But the gate already knows I don’t feel like saying goodbye?” The story did not answer. She answered her own question eventually.
That is the practical part that matters. A personalized story is not a magic trick. It works when you treat it as a conversation starter, not a replacement for your voice. Here is what I did, and what I would suggest to another parent: I read the story out loud, but I stopped at the gate. I asked her what she thought the gate was for. That pause was the whole point. She got to steer us back into the story. With a regular book, a child knows the plot is fixed. With a story built around their name and a real daily struggle, the plot feels flexible. They can try out a line and see if it holds.
Dan told me he made one change before inviting other families in. He added a rule to the system: every generated story has to include a small mistake made by the hero. Not a character flaw, exactly, but a mess. “If your kid only ever wins,” he said, “then failure starts to feel like a bug in the universe. So I ask for a part where the hero drops something, breaks something, or refuses to listen once.” That is the kind of detail many personalized products skip. Let the hero forget the map. Let the hero blame the elephant. Seeing your own name next to a small failure, while a calm narrator keeps going, is more grounding than a hundred motivational posters.
There is also a lower-tech version of this, and I do not say that to dismiss his tool. You can do it with a sticky note. Write your child’s name over the word “Riley” in a book. Ask a question at the end of a page and let your child answer as themselves. What matters is not the technology. It is the invitation to be the protagonist, which the machine simply makes fast. I would caution against handing a kid a screen and treating the story as a babysitter, the same way I would caution against any screen at bedtime. The glow is still the glow. The value is in the shared reading, a physical shoulder against a physical shoulder.
On the third night, my daughter requested “the gate story” again, but she wanted me to read it with a different ending, one she invented herself. That was a better result than any computer could generate. The tool gave her the scaffold; she gave it the roof.
Dan is still fiddling with his beta, adding a place where the parent can leave a comment for the next night’s story, the way you write “to be continued” on a lemonade stand sign. He is not pretending it will fix bedtime. “Half the nights my kids fall asleep before I finish,” he shrugged. “The point is they fall asleep here, in the story, not staring at a ceiling reliving whatever bothered them.”
I am still not sure how much of this I want in our normal routine, or where the line sits between a useful prompt and polished toddler marketing. But the next time my daughter tells me she is not sleepy enough yet, I suspect I will type a short prompt about a lost water bottle, a grumpy owl, and the slow walk home through wet grass. I will stop around page three, and I will wait for her to decide. That is not the future of parenting. It is just Tuesday night, with a little help from a dad who was tired of asking his kids to choose a book they did not care about.
Please indicate: Thinking In Educating » Why a father of three built an AI that writes bedtime stories where your kids are the heroes