BRAHMASTRA

BRAHMASTRA is the ultimate on device study superpower for Arm powered mobiles a privacy first, offline AI study suite with games, quizzes, mindmaps, audio, YouTube curation, and universal navigation.

Comment

Inspiration

BRAHMASTRA started from a simple question: “What if a phone could be a complete, offline study companion powered entirely by on-device AI?” Seeing the Arm AI Developer Challenge push for efficient, privacy-preserving AI on Arm devices inspired the idea of combining quizzes, mind maps, games, and summaries into one cohesive app. The name “BRAHMASTRA” reflects the goal: a single, powerful tool students can rely on anywhere, even without internet access.

What it does

BRAHMASTRA lets users snap a photo of notes or textbooks and instantly generates quizzes, mind maps, audio summaries, and even study games from that content, all running locally on an Arm-powered device. It also includes a context-aware AI chatbot, YouTube learning recommendations, a concentration game builder, and a clean dashboard that tracks progress over time.

How we built it

Using NPL and Google AI Studio, you refine and test AI prompts and pipelines. Loveable AI and Tempo accelerate AI model integration and UI workflow development, enabling rapid prototyping while keeping close to arm-optimized inference. VS Code provides the environment for custom Android app coding, performance tuning, and final packaging on Arm64 devices.

Challenges we ran into

The biggest challenge was fitting a useful multimodal model on-device while keeping latency and memory usage low enough for a smooth mobile experience. Designing a UI that exposes many powerful features (study tools, games, chatbot, dashboard) without overwhelming users required several iterations on navigation, theming, and accessibility.

Accomplishments that we're proud of

BRAHMASTRA delivers a full pipeline—camera to quiz, mind map, audio, games, and analytics—without relying on any cloud inference, which showcases what Arm-based devices can do at the edge. The concentration game builder that turns a user’s own notes into memory and focus games is a unique feature that makes studying feel engaging rather than repetitive.

What we learned

Building this app highlighted how critical model choice, quantization, and hardware-aware optimization are when deploying AI to mobile devices. It also reinforced the value of thoughtful UX: clear navigation, dark/light themes, and simple flows are just as important as the underlying AI when targeting real students and developers.

What's next for BRAHMASTRA

Next steps include adding collaborative features so study groups can share quizzes and games, plus adaptive learning paths that adjust difficulty based on ongoing performance. There is also room to support more languages, expand to iOS with similar Arm optimizations, and integrate smaller specialized models for math, code, and science explanations to further improve accuracy and speed.

Built With

  • ai-studio-platform-for-deployment-and-hosting
  • browser-based-localstorage-for-state-management
  • component-based-architecture-with-typescript-interfaces-for-type-safety
  • css
  • css3-animations-and-keyframes-for-splash-screen-and-ui-transitions
  • esm-module-system-with-import-maps
  • google-generative-ai-sdk-1.30.0-(gemini-2.5-flash-model)-for-ai-powered-puzzle-generation-and-chat-responses
  • html
  • javascript
  • lucide-react-0.554.0-for-icons
  • node.js-runtime-environment
  • react-19.2.0
  • react-hooks-(usestate
  • recharts-3.4.1-for-data-visualization
  • responsive-web-design-with-mobile-first-approach
  • rest-api-integration-with-cors-proxy-configuration-through-vite-dev-server
  • space-grotesk-font-from-google-fonts
  • tailwind-css-(via-cdn)
  • typescript-5.8.2
  • useeffect
  • useref)-for-state-and-lifecycle-management
Share this project:

Updates