Forecasts weekly US natural gas storage from the weather, then scores itself against the analyst consensus.
- Python
- LightGBM
- GitHub Actions
Jake Adler
Ex-Founder. AI Builder.
Technology feels most magical when you can hold it.
I spent a year building an AI children's book company as a solopreneur, then walked away. This is the whole story.
Read the story
The founder years
StorybookYou: an AI-personalized children's book company
- ~$0
- in revenue
- 0+
- storybooks printed
- 0+
- customers
Chapter 1
The idea
What if you could place kids in printed children's book with just one photo? In April 2025, image models were just getting good enough to make this possible. I made the first books for friends, using a full AI pipeline for both the writing and the illustrations.
Photo: early prototypes
Chapter 2
Zero to $70K
AI-writing still outputted too much slop. And, I realized the real excitement — and the scalability — was in photo personalization. So I pivoted to templated, human-written stories starring each kid.
Photo: Cumulative revenue, Stripe
$70K across 1,758 orders.
11 titles, oldest to newest
Template, then your child.
A Royal Night at the Castle
Feb 2026
The Great Easter Bunny Rescue
Mar 2026
As Far as Love Can Go
Mar 2026
Revolutionary War Adventure
Apr 2026
Big Hands, Little Hands
May 2026
Builds a Road
Jul 2026
The Daytime Stars
Aug 2026
The Mac 'n Cheese Mission
Aug 2026
Skeleton Sleepover
Aug 2026
The Guiding Light of Diwali
Sep 2026
The Lights of Hanukkah
Oct 2026
Chapter 3
The Diwali spike
Our Diwali storybook was the first ever human-illustrated, AI-personalized children's book. Brought to life with an Indian artist, it brought in about $40K in six weeks.
Photo: Daily sales, Sep to Oct 2026
$40K in six weeks.
Chapter 4
Scaling as one person
I couldn't review every book myself, so I built a bespoke review platform and hired a small team. Review time dropped from ten minutes to ninety seconds per book.
Review time after the custom review platform
Review time per order
Before
10 min
After
90 sec
6.7× faster
Chapter 5
Where every dollar went
The honest version of the numbers. Seasonality, low repeat purchases, and unpredictable image output kept margins thin without an end in sight, so I decided to stop.
All time, August 2025 to today, per-dollar breakdown
The honest reasons I closed it: low repeat purchase, strong seasonality, and image generation that could never be fully predictable.
Chapter 6
What I took with me
I learned what it takes to build a business and how to design a product alongside customers. I also learned that distribution and sales matter as much as the product, perhaps even more so.
Photos: TomTom pitch night, podcast, happy customers
The writer
What I think about when I'm not building.
Side projects
Every promo code buried in YouTube video descriptions, in one searchable place.
- TypeScript
- Python
- Supabase
A stock ticker with pixel-art animation, because markets should be fun to watch.
- TypeScript
- VISX
- Finnhub
A Chrome extension that shows clothes on you, powered by Nano Banana.
- Python
- Chrome
- Nano Banana
Incoming call
JJarvis
Your 7:00 AM briefing
An AI secretary that phones you the news when you wake up.
- Python
- Twilio
- ElevenLabs
An early attempt at an AI tutor for high schoolers, with teacher oversight.
- JavaScript
- GPT-3.5
- Firebase
Who is Jake?
An avid traveler, dog lover, and adventurer.
Traveler
South America, mostly. I love using my limited Spanish.
Bartender
Where I learned to be customer-first.
Dog-lover
This is Django. Django is the best.
Scuba
My favorite activity on Earth.
Say hello.
Building something physical, something AI, or something that needs both? I'd like to hear about it.
© 2026 Jake Adler · New York