Skip to content

Case study · 02 / 04 · 2026

An independent product where families turn memories into illustrated storybooks. The generation pipeline produces a book preview in about 30 seconds end to end.

Role

Founder & engineer

Timeline

2025 → 2026

Stack

Next.js · TypeScript · Medusa · AI story & image generation · Stripe

Little Legend cover

Context

Parents want a keepsake, not another app. Little Legend turns a few family details and memories into a personalised, illustrated storybook: preview today, PDF tonight, hardcover for years.

What I built

  • The full product flow: memory capture, story structure, generation hand-off and a polished storefront on Next.js.
  • Commerce and order logic in Medusa, so books, previews and hardcovers are real products with real checkout.
  • An AI story-and-image generation pipeline with structured prompts and review steps; end-to-end story generation runs in about 30 seconds.

Architecture

Next.js owns presentation and the personalisation wizard. Medusa owns catalogue, cart and orders. The generation pipeline is a separate, queue-driven step so a slow model never blocks checkout, and every generated asset is reviewed and stored before it is shown to a family.

Outcome

An independent product, live at littlelegend.online, built and shipped solo. It is the reference for how I approach AI features: as product plumbing with limits and review, not as a demo.

Screens

  • Little Legend book catalogue

  • Little Legend book page, ready to personalise

Next project