From a parent's idea to an AI-powered storytelling application.

GiMagi began with a simple observation from parenthood: children don't always want to hear someone else's story. They want to help decide what happens next.

I saw an opportunity to create a storytelling experience that makes room for a child’s imagination while giving parents a straightforward way to create engaging stories for different moments throughout the day—from imaginative adventures to quieter bedtime experiences. The guiding principle became: “GiMagi doesn’t tell children what to imagine. GiMagi brings their imaginations to life.”

I developed the core product experience around the GiMagi 5—Poppy, Milo, Beep, Dotty and Ember—and five distinct storytelling worlds, with child-led choices shaping each adventure. The story journey includes choosing a character, exploring a setting and selecting a storytelling style.

Rather than relying on a single level of complexity, I designed three storytelling styles—Wisher, Dreamer and Wonderer—to offer different levels of language, detail and adventure. The experience supports both Anytime and Bedtime storytelling, allowing families to choose between imaginative adventures and gentler stories designed for winding down.

I took the idea beyond concept and into a live, working alpha. Using AI-assisted development tools and a workflow built around GitHub, Cloudflare Workers and Gemini, I coordinated the experience from product decisions and content direction through implementation, deployment and troubleshooting.

My contribution is product ownership and execution: setting requirements, making design and experience decisions, integrating components, evaluating output and iterating. AI assists with development and story generation; this is not a claim that I personally hand-coded every component.

The result is a functioning storytelling journey with original characters, world selection, customizable story styles and generated stories that families can read.

These observations have reinforced the importance of recognizable characters, engaging visuals and child-led exploration. Broader testing with parents and children is the next step, focusing on usability, storytelling quality and engagement before commercial launch.

Early observations and next steps. Early family testing has provided encouraging feedback on GiMagi’s characters and visual experience. My daughter has begun recognizing characters by name and independently asking where they can explore next—an early indication that the characters and worlds are memorable and encourage curiosity.

To address this, I developed detailed Character Bibles for the five founding characters, defining their personalities, motivations, characteristic behaviors and boundaries. Initially incorporated into the story-generation process, these evolved into reference documents as the engine developed. They provided a consistent standard against which I could evaluate generated stories, identify character inconsistencies and guide further refinements.

Creating consistency through Character Bibles. One of the earliest challenges was ensuring that GiMagi’s characters remained recognizable across different stories, settings and situations. AI-generated narratives could be imaginative and engaging while still drifting away from an established character’s personality.

Reaching a working alpha was a milestone, not the finish line. I have generated approximately 250 stories during development and quality testing (estimated), comparing iterations of the story engine, including V2 and V3 approaches, to assess narrative structure, age-appropriate language, character consistency and variety.

Alongside development, I have explored the commercial model: researching competing products, evaluating potential subscription tiers, considering operating costs and mapping a phased path toward launch.

These are planning assumptions and experiments, not proven market demand or realized revenue. GiMagi remains in active development, with further product refinement and validation ahead.