Lil' Mimic Studio
aikidsreactgemini
October 2025 Live

Lil' Mimic Studio

A dad-built AI character creator for toddlers. Upload a photo, generate a character, make a comic strip. Powered by Google Gemini. 10,000+ generations and counting.

↗ View Live

Built with

React
React
Gemini
Gemini

Why I Built This

My son Dean was 18 months old and obsessed with cartoons. I wanted to make him the character — not some generic avatar, but a recognizable version of his actual face drawn in the style of whatever show he was watching that week.

Standard AI image tools weren’t built for this. Too many steps, too much prompt engineering, no good story-to-strip pipeline. So I built one.

What It Does

Lil’ Mimic Studio is a React web app powered by Google Gemini that turns a photo of any child into:

  1. Custom Character — AI generates a cartoon illustration in the requested style (superhero, anime, storybook, etc.)
  2. Character Sheet — Name, personality, powers, backstory — all hallucinated from the photo
  3. Comic Strip — A 3-panel story featuring the character in an original adventure

The whole flow takes under 60 seconds. Upload a photo, pick a style, watch the magic.

01
Photo Upload
Parent uploads child's photo + picks a style (superhero, anime, storybook…)
02
Character Gen
Gemini analyzes face features → Imagen renders cartoon illustration
03
Character Sheet
Name · powers · backstory hallucinated from the image
04
Comic Strip
3-panel original story featuring the character — shareable in under 60s

Technical Stack

  • React + Vite — Frontend, hosted on Google AI Studio
  • Gemini 1.5 Flash — Image understanding + character generation
  • Gemini Imagen — Illustration rendering
  • Google AI Studio — Hosting platform (no backend required)

The Numbers

  • 10,000+ generations since launch
  • Shared across parenting communities without any paid distribution
  • Built in a single weekend session during Dean’s nap

Lessons

Gemini’s multimodal understanding is genuinely impressive for face-aware generation — it picks up on specific features (hair color, eye shape, cheek roundness) and preserves them through the cartoon transformation. The hardest part wasn’t the AI — it was the 3-panel narrative prompt that produces coherent, age-appropriate stories without supervision.

The real win: Dean now asks to “make a comic” as a regular activity. That’s the metric that matters.

← All Projects