top of page
Covercub_Strip.png

CoverCub


AI-assisted personalised children’s books (solo build)

The Product
A web platform allowing parents to create custom-illustrated children’s books using Generative AI, with a focus on high-quality storytelling and design.

Link to the site: https://covercub.com/

CoverCub started from something very simple:
I wanted to create a real children’s book for my 2-year-old son.
I tried the existing “personalised book” websites and nothing felt right.
The stories were flat (“my daughter is a princess”, “my son is a superhero”), the illustration style was often very Disney-like and over-polished, and the whole experience felt more like a gimmick than a meaningful object you’d keep on a shelf.
So instead of forcing myself to like those products, I decided to build the version I wished existed, and use it as a playground to combine my skills as a designer and creative with the new AI tools I was learning and my (very) old coding skills that I brought back to life.
Screenshot 2025-12-11 at 11.35.37.png

The concept
CoverCub was a web app where parents could choose from a curated library of stories I wrote myself, and turn them into a personalised, illustrated book where their child becomes the main character.
The parent picked a story theme (bedtime, courage, first day at kindergarten, etc.). Uploaded a photo of their child. Added name and a few simple details. Chose an illustration style.
Behind the scenes, I used AI to generate the illustrations based on my pre-written stories and page templates, so the storytelling quality stayed high and intentional, while the visuals and character were adapted to the specific child.
I wasn’t trying to “invent” personalised books. I was trying to do them better: deeper stories, more tasteful illustration, and a smoother product experience.

Screenshot 2025-12-11 at 12.32.28.png

From idea to working product
A true solo project – I was responsible for:

  • Concept, product strategy and brand identity.

  • UX / UI in Figma and overall creative direction.

  • Frontend and backend build.

  • AI prompt design and integrations.

  • Payment, printing pipeline and basic marketing.


I started with Figma flows and the brand, then:

  • Built the first MVP with low-code tools (like Lovable), then moved to Cursor when I hit their limits.

  • Re-learnt frontend and backend and rebuilt it as a proper web app.

  • Used Supabase for user accounts and secure data.

  • Connected to AI services for image generation based on templates and uploaded photos.

  • Generated print-ready PDFs and integrated with a print house.

  • Set up checkout, transactional emails and simple Meta ads using a mix of AI visuals and Adobe tools.
     

Screenshot 2025-12-07 at 20.03.45.png

Outcome and lessons

Shortly after I had a working version, Google released Storybook, a free tool where parents can upload a few photos, write a sentence about the story, and get a personalised children’s book in minutes.
That changed the landscape: my paid, bootstrapped product was suddenly competing with a free, “good enough” solution from Google.
Instead of trying to fight that head-on, I made a conscious decision to shut CoverCub down and move on to my next idea.
 
For me, CoverCub wasn’t a failure; it was a complete learning loop:
  • Taking a personal frustration and turning it into a product.
  • Building a full stack (UX, brand, frontend, backend, AI, infra, checkout, printing) on my own.
  • Understanding how fast AI-driven markets can shift, and when it’s smarter to kill a product and apply the skills to the next one.
Marketing & AI Video Production
Go-to-Market Strategy I treated the project as a full DTC brand launch. Instead of hiring a creative studio, I engineered a custom Generative Video Pipeline to produce high-end ads for Meta. I used Nano Banana Pro to generate consistent character assets, animated them with Google Veo 3 and Kling AI 2.6, and finalized audio with ElevenLabs, allowing for rapid creative iteration at zero production cost.
bottom of page