MindFlowwAI

Plans list
MindMap canvas
MindMap canvas
Metal shader loading animation
Plan generation form with Gemini
Gendered plan short preview
Tasks overview

Inspiration

MindFlow started from the simple idea that if you can describe a goal in words, an AI should be able to turn it into a Mind Map, not just a checklist. We wanted to move beyond chat‑style AI and explore how Gemini could act like a “planning engine” inside a productivity app.

What it does

An app for creating and storing plans in the form of a roadmap. Users organize tasks, establish connections between them, and clearly see the dependencies of each step, all displayed in a clear, visual Mind map layout. You type a goal, import PDFs, notes, or images, and Gemini turns them into structured tasks on the same canvas. All your ideas and plans in your pocket.

How we built it

MindFlow uses Swift & SwiftUI, SwiftData for local storage, Firebase AI Logic SDK for plans generation, Metal shaders for smooth canvas interactions, and custom generic tree data structure for Mind Map representation. Each request is a compact, schema‑aware payload; Gemini returns a structured plan that the app maps directly to the UI.

Challenges we ran into

Mind Map canvas zoomable and draggable layout in SwiftUI
Convenient FirebaseAI Gemini integration with custom Gemini Service
Complex interactive Metal shader animation for loading state
MVVM architecture
Generic tree custom data structure for tree based Mind map
SwiftData persistence

Accomplishments that we're proud of

Pioneered a complex, native mind map interface in SwiftUI that delivers a fluid, zoomable, and draggable canvas for visualizing plans. Integrated Google's Gemini AI natively to generate structured, actionable plans directly from simple text goals. This makes MindFlow the first SwiftUI application to combine real-time AI-powered planning with an intuitive, fully interactive mind map experience.

What we learned

We dove deep into the Firebase AI SDK for iOS, learning to configure the Gemini model and stream real-time AI responses directly within SwiftUI. This project gave us hands-on experience structuring prompts, managing model outputs, and integrating advanced AI natively into a fluid, interactive mind map interface.

What's next for MindFlow

We’ll introduce shared plans and templates for teams, deeper progress tracking, and time‑aware plan adjustments for busy periods, while expanding to iPad with seamless sync. We’ll continue exploring Firebase AI and other Google tools for developers, integrating new Gemini features to make planning smarter and more adaptive.