QFlash

• 

  logo

• 

  deck screen

• 

  chat screen

• 

  make card screen

• 

  study screen

Inspiration

• Wanted a faster, more joyful way to memorize complex topics on mobile.
• Frustrated by rigid flashcard apps; inspired to blend AI-assisted authoring with proven SM‑2 spaced repetition.
• Aimed for a delightful SwiftUI experience with smooth animations and on-device feel.

What it does

• Create/manage decks with icons/colors; add text or image-backed cards.
• AI chat to generate cards from your prompts or conversation.
• Study with SM‑2: automatic scheduling, difficulty tracking, due queues.
• Image tools: generate/crop images, upload to Supabase Storage, serve via signed URLs.
• Share/import decks via deep links with short URLs.
• Fast, offline-first UX with lightweight persistence.

How we built it

• Stack: SwiftUI + MVVM, Supabase Edge Functions (Deno/TypeScript), Supabase Storage.
• Architecture: Models hold business logic; ViewModels manage UI state and Bindings [[memory:4607492]].

Challenges we ran into

• Reliable token streaming and UI synchronization on iOS.
• Designing prompts that yield concise, high‑quality cards.
• Image pipeline: client resize/crop → upload → signed URL caching.
• Balancing animation polish with list/grid performance.
• Handling edge cases in SM‑2 (resets, lapses, clamped EF).

Accomplishments that we're proud of

• Cohesive study flow with smooth SwiftUI transitions.
• One‑tap “turn chat into cards” workflow that actually saves time.
• Robust, minimal Supabase integration (Edge Functions + Storage) without heavy SDK overhead.
• Clear separation of concerns that made testing and iteration fast.

What we learned

• Small UX details (typing/streaming indicators, safe array bindings) massively improve perceived speed.
• Prompt engineering for structured outputs is as important as model choice.
• Supabase Edge Functions are great for secure, server‑side AI calls and media tasks.
• SM‑2 tuning benefits from real usage data; defaults are just a starting point.

What’s next for QFlash

• Adaptive SM‑2 tuning from user performance data.
• Camera OCR to turn photos into cards; speech playback for hands‑free review.
• Collaborative/shared decks and public gallery.
• Widgets and Live Activities for bite‑size reviews.
• Optional cloud sync and richer analytics with privacy controls.

Built With

• deno
• ios
• ios-file-system-(documents)
• mvvm
• sm-2-spaced-repetition
• supabase
• supabase-cli
• supabase-edge-functions
• supabase-storage
• swift
• swiftui
• typescript