Why This Dad Built an AI to Write Bedtime Stories Where Your Kid Is the Hero
The baby had finally stopped crying, but my middle kid, age four, was standing on her bed doing a slow-motion jump that she called “statue jumping” and my oldest, seven, was lying sideways across his pillow, arms crossed, announcing that he would not sleep unless the story was about him. Not about a boy like him. Not about a kid who liked dinosaurs. About him, specifically. By name.
I knew the feeling. After a long day, the last thing you want to do is invent a fresh, engaging narrative about a child named Theo who wields a toothbrush sword and rescues his stuffed octopus from the vacuum cleaner. I usually default to a classic fairy tale or an old standby, but Theo was not having it. He wanted to be in the story. Not as a background character. The hero.
So my wife said, “You should just build something for that.” And because I’m a software engineer and also too stubborn to accept that I could just buy a book with blank-name stickers, I did.
The result is a small tool that takes a few details — your child’s name, a favorite toy, a recent worry or a silly scenario — and writes a short bedtime story around them. It’s been live for a couple of weeks as a free beta. But what surprised me wasn’t the technology. It was watching my oldest kid sit up straighter when he heard his own name in the first sentence.
The first story the AI wrote was simple: Theo finds a lost key that unlocks a door in the backyard hedge, and behind it is a small, grumpy badger who needs help finding his glasses. The writing was not Pulitzer-level. The badger sounded suspiciously like a grumpy dad. But Theo was hooked. He asked me to read it twice. Then he asked if I could change the badger into a turtle.
I changed it in the prompt and we got a turtle with the same problem. Theo was delighted. It took me about ninety seconds. My wife raised an eyebrow. I pretended I meant to teach him about iteration.
The first time we tested it, I thought the point was just the novelty. Hearing your own name in a story is a thrill. It’s the same reason parents doodle their kid’s name on a backpack label and kids spend the first week of school showing it to everyone. It makes something general feel specific. But the more we used it, the more I realized what was actually happening: Theo wasn’t just listening. He was problem-solving.
For a few nights, I read a new story where he had to convince a stubborn robot to let him through a gate. He didn’t wait for me to tell him what happened. He started arguing with the robot from his bed. “We could just go around the gate!” he shouted. I said, “The story says the robot has a force field.” He said, “Then we dig a hole.” That’s not in the story. But it’s better.
That’s the thing about making your child the protagonist. It flips a switch from passive listening to active engagement. They’re not watching a character solve a problem; they’re being asked how they would solve it. And that’s a skill that transfers to real life — calming down at bedtime, trying a new food, or getting through a tough moment at school. A story is just a low-stakes rehearsal for the kind of thinking you want a kid to do the next day.
I don’t think you need an AI to do this, honestly. I’ve been doing it more by hand now, too. The trick is to take a familiar story and swap in your child’s name and something specific from your house. Instead of “Once upon a time, there was a little girl named Goldilocks,” try “Once there was a girl named Maya who walked into a house where the chairs were just her size and the cereal bowls were still warm.” Then let the story go sideways. You don’t need a plot. Kids will supply one if you give them a problem. Ask, “What does she do when she hears a big stomping sound outside?” and wait. You’ll get a better answer than any script.
One thing I’ve learned is that the AI makes the child a little too perfect. In an early version, the hero always solved the problem on the first try, was brave, and remembered to share. That’s nice, but it’s not true. And kids know. Theo started to lose interest in the stories where he never struggled. So I changed the prompt. Now every story has a mistake in the middle. The hero forgets something, or says something unkind, or gets scared. Then they have to fix it. That feels more like real life, and it makes the ending matter more. I’ve started doing this with the stories I tell out loud too. Instead of making him the kid who always wins, I make him the kid who breaks something, feels bad about it, and finds a way to make it right. He listens more carefully when the character is flawed. I think it gives him permission to be messy.
The other night, the AI generated a story where the hero — a boy named Theo with a red dinosaur backpack — was afraid of the dark. That wasn’t something I typed in. The AI must have inferred it from a previous prompt where I mentioned his nightlight. Theo stopped me mid-sentence. “Why is he scared of the dark?” I said, “I don’t know, why do you think?” He thought for a second and said, “Because sometimes you can’t see if there’s a monster, but the light helps you see that it’s just a chair.” Then he turned over and fell asleep. That conversation was worth more than the story.
I’m not sure this tool will change bedtime permanently. The beta runs out in two weeks, and I’ll probably keep tinkering with it. But what it gave our family isn’t just a novelty. It’s reminded me that a story doesn’t need to be polished or elaborate. It just needs to make the child feel like the narrator is talking to them, not at them. You can do that with a name, a small problem, and a chance to ask “What now?” and actually wait for the answer. The AI is just an easier way to get there when you’re worn out and your brain is blank.
Tonight, I’ll probably let the tool write the first draft, then I’ll change the badger to a turtle and let Theo interrupt.
Please indicate: Thinking In Educating » Why This Dad Built an AI to Write Bedtime Stories Where Your Kid Is the Hero