Note
This project is intended to allow children who can't read/write (~5 years old) to create picture books by simply recording a story idea. Try it: https://josefs-picture-book.brandstetter.io
I've asked my 5-year-old son what he would like the computer to do for him (he uses ChatGPT quite a lot at home).
He said: "I want the computer to create a picture book and read it to me."
Since I really enjoy all of the new AI tools, I spent a couple of evenings creating this project for him.
- v0: I used v0 for most of the UI. I even used the shadcn CLI to add components.
- Cursor: I used Cursor mostly for the same things I've used GitHub Copilot for, primarily to fix some CSS and to scaffold functions.
- Whisper: Whisper is a great resource to transcribe text. What's awesome is it can deal with many languages (we speak English, German, and Arabic at home).
- GPT-4o: I used GPT-4o to generate a main story and to generate DALL-E prompts for the individual pages. I really love the structured output (hint: use Zod to stay type-safe all the way).
- DALL-E: I used DALL-E to generate the images. Sadly, it doesn't have a way to keep the characters consistent. I might use some other GenAI tool in the future.
- Copy the
.env.example
file and name it.env.local
. - Install packages:
npm i
. - Start the server:
npm run dev
.