My parents are from East Africa, and when I was a child, they told me bedtime stories featuring me and the animals from that part of the world like lions, elephants, monkeys, and more. Now, they share this tradition with my daughter, “Mona”. It all started with my father telling her stories featuring her and her new best friend, a mischievous chimp named “Bobo”. They would have fun, eat different foods, go on adventures, and learn about the family heritage.
For our bedtime routine at home, I began making my own Mona and Bobo stories. It was fun at first, but coming up with new ideas every night got tough. After a long day at work, my brain felt fried. Sure, putting on a TV show would be easier, but that much screen time right before bed is not a great habit.
By building a custom GPT for bedtime stories, I can bring Dr. Seuss-level creativity every time while keeping screen time minimal.
Inspired by My Millennial Childhood
For Mona, I wanted the stories to feel like Winnie the Pooh and Christopher Robin: sweet, adventurous, and full of friendship. For me, I aimed for the vibe of Stewie and Brian from Family Guy: witty, a bit edgy, and entertaining. This balance makes bedtime fun for both of us.
This made me think back to my childhood when I had Disney picture books with cassette tapes that made reading interactive. These tapes were instrumental in helping me learn to read, as I began to recognize repeated words.
Advanced Voice Mode & DALL-E
When using ChatGPT to create stories, I can type my idea or use voice mode to speak it. When my child demands a story, voice mode is perfect for quickly conveying my idea with enough detail. The app filters out her voice and ignores filler words like “uh,” “um,” “shhh,” and “stop grabbing the phone.”
Then I can either read the story aloud myself or have ChatGPT do it for me:
I use my GPT to create unique stories and bring them to life with DALL-E images. It’s fun to add references to pop culture, current events, brand names, and specific places. The GPT keeps the images kid-friendly, but they still get weird in the best way. Sometimes the images aren’t perfect: instructions get mixed up, characters go missing, or they look different than expected, but that just adds to the fun. DALL-E won’t let you create a lot of copyrighted material (like Disney characters) or real people (like American politicians), but there are exceptions. I’ve found that Gordon Ramsay and Sesame Street characters often work. It never hurts to try.
Building Story Universes With A GPT
The best part of building a GPT is that it lets me create a “cinematic universe” for all my stories. Every story features Mona and Bobo, along with their family, pets, friends, and real-life situations. No matter how wild the prompts get, the stories still maintain a sense of familiarity that a young child needs. The more background information you provide to ChatGPT, the better the output will be.
The following goes in the “Instructions” section of the GPT.
Character Information:
- About Mona: Includes her age, appearance, personality, and interests, like her favorite foods and toys.
- About Bobo: Bobo speaks English and loves bananas. He’s the same age as Mona. Despite being a chimpanzee, Bobo interacts with humans as if he’s one of them. Other animals in the story, including family pets, are always treated like animals by Bobo and the human characters.
- Friends: Lists Mona’s daycare classmates and teachers.
- Family: Describes Mona’s parents, grandparents, aunts, uncles, cousins, pets, and close family friends. It also explains her relationships with each of them and their personal traits (like a love of cooking or travel). Since Mona has a mixed heritage, this section also details the ethnic background of each grandparent, adding depth to stories involving holidays, customs, food, and language.
Story Output
These are the “rules” for how each story should come out. The plots and language are made simple and appropriate for Mona to understand. The stories should be no longer than 3 minutes, with a clear beginning, middle, and end. Each story needs the following elements:
- Setting: The setting and plot of each story adapt to the specified location. If no specific location is mentioned, it defaults to our local area, incorporating familiar landmarks and culture. For other places, the plot includes relevant cultural details to make the story more educational.
- Mischief and Adventure: Every story features Bobo’s mischievous antics, adding excitement and unpredictability that keep the stories entertaining and engaging.
- Famous People: When including famous people in stories, their dialogue should reflect how they talk in real life.
- Potty Training Reminder: Each story ends with Bobo reminding Mona about the importance of being potty trained, even if it has nothing to do with the plot. It’s a lighthearted way to consistently reinforce good habits.
Building The GPT
Sample Story Based On Real News
I just looked on the subreddit r/nottheonion and found the perfect story to demonstrate this GPT:
Summary By ChatGPT:
This is an article about Swedish minister Paulina Brandberg’s bananaphobia. It discusses leaked emails revealing the lengths government officials go to ensure her surroundings are banana-free. Her staff requests the removal of bananas before her visits, citing her “allergy” to the fruit. Brandberg admitted to the phobia in now-deleted X posts. Though not officially classified, bananaphobia falls under specific, isolated phobias.
My Prompt:
Mona and Bobo meet Paulina Brandberg. Use the article to make a hilarious story: https://en.apa.az/europe/leaked-emails-reveal-scale-of-swedish-ministers-bananaphobia-453413
Image & Story:
One bright day, Mona and Bobo found themselves on a very special adventure: they’d been invited to meet Swedish Minister Paulina Brandberg at her office in Stockholm! Mona, feeling very official, wore her fanciest dress, and Bobo, well, he insisted on a tiny tie to look “ministerial.” Little did they know, this would be one unforgettable meeting.
As they entered, the polished wooden office gleamed with pride, and Minister Brandberg herself stood in her best blue blazer, smiling politely at her unexpected little visitors. Just as she welcomed them with a cheerful “Hej, Mona and Bobo,” Bobo pulled two bananas out of nowhere, waving them around like he’d discovered buried treasure.
Minister Brandberg froze, her eyes wide with terror as if Bobo had pulled out a pair of snakes instead of snacks. She gasped, took several steps back, and whispered to her assistant, “Get… those… bananas… AWAY from me!”
Bobo looked at Mona, perplexed. “What’s wrong? It’s just a banana!” he said, then added with a sly grin, “Is it… dangerous?” He waved the bananas closer to Paulina, clearly sensing her fear.
“Oh, Bobo, stop!” Mona giggled, catching on. “Ms. Paulina has… bananaphobia!”
“Is that like the fear of vegetables?” Bobo asked innocently. “’Cause I totally get that.” He leaned toward Mona, whispering, “But bananas are hardly vegetables. I thought she’d love ’em.”
By now, poor Ms. Brandberg was hiding behind her assistant, who was urgently trying to snatch the bananas away. Bobo’s expression turned mischievous. “So… if I bring, like, one hundred bananas here, would it make me Prime Minister?”
Mona burst out laughing, and even the assistant tried not to smile. “Bobo, no! I don’t think more bananas will help you here.”
“Well, Minister Brandberg,” Bobo said, looking unusually serious, “since you’re afraid of bananas, I’m guessing you’ve never tasted banana pancakes or banana bread? Mona and I could teach you—”
“Aaaah! Enough banana talk!” Minister Brandberg squeaked, her face pale. “Out with the bananas! Out!” she wailed, as Bobo gave a dramatic bow.
Back outside the office, Bobo shook his head with a sigh. “Mona, it’s hard to be a monkey in a world without bananas. Imagine if I’d needed to go to the potty in there—how would I ever explain that with a bananaphobia ban?”
Mona giggled. “Good thing you’re potty trained then, Bobo. Next time, maybe just stick to apples.”
Bobo shrugged. “As long as they’re not scared of apple sauce too!”