Parents want engaging and personalized content for their children, but most story apps are generic and one-size-fits-all. They cannot reflect a child’s unique interests, appearance, or family dynamics. Creating fully custom stories with illustrations and narration traditionally requires significant time, artistic skill, and expense, putting it out of reach for most families.

As the Lead AI & Full-Stack Developer, I was responsible for the entire architecture and execution of the application. My role included:

Designing and engineering the backend orchestration layer for Google’s generative AI (Gemini)

Developing complex TypeScript Cloud Functions to manage the multi-step generative process (text, image, audio)

Building the interactive frontend in FlutterFlow for an intuitive user experience

The project’s goal was to put the power of a creative studio into the hands of parents. Key objectives:

Deep Personalization: Enable parents to control characters, plot, visual style, and narration for a deeply personal story

Automated Multi-Modal Creation: Seamlessly generate and combine text, images, and audio into a single interactive storybook

Technical: Build a scalable application capable of handling complex API chains, managing generated assets, and delivering them to a custom story player

7akai transforms storytelling into a collaborative, infinitely creative experience:

Unprecedented Personalization: Children see themselves and their world reflected in stories, boosting engagement and literacy

Empowers Parents: Allows parents without artistic skills to become creators, enhancing bonding experiences

Demonstrates Multi-Modal AI: Showcases advanced prompt engineering, API orchestration, and polished consumer-facing AI applications

Limitless Content Library: Generates stories on demand, providing endless entertainment without traditional production costs

The application was built on a modern tech stack designed for multi-modal generative AI workloads.

Core logic implemented in TypeScript Cloud Functions to orchestrate calls to Google AI’s Gemini models

Workflow:

Ensures consistency across text, images, and audio

Keywords: Google AI, Gemini API, Prompt Engineering, Multi-modal Generation, Generative AI, TypeScript, Cloud Functions, Vertex AI

Interactive, user-friendly interface to define story parameters step-by-step

Custom "storybook player" widget assembles AI-generated text, images, and audio dynamically

Editing interface allows parents to arrange elements on each page

Keywords: FlutterFlow, Interactive UI, Dynamic Content Assembly, Custom Widgets

Firestore: Stores user accounts, saved story configurations, and metadata pointing to generated assets

Firebase Authentication: Secure account management for parents

Firebase Storage: Stores final AI-generated image and audio files for fast, reliable retrieval

Keywords: Firebase, Firestore, Serverless, Data Management, Firebase Storage