Aura Vega TV — Living Room Consensus Engine

Eliminate the 20-minute "what should we watch?" decision fatigue on Fire TV with an explainable multi-viewer consensus engine and instant playback.

Comment

Inspiration

Every evening, millions of living rooms face the dreaded "20-minute stalemate" — couples, roommates, and families endlessly scrolling through fragmented streaming apps before settling on background noise or giving up in frustration.

The living room television is the ultimate shared household screen, yet modern streaming interfaces treat it as nothing more than an oversized single-user smartphone on a wall. We set out to change that. We were inspired to build a native Fire TV experience on Amazon's next-generation Vega OS that respects the shared nature of the living room: serving as an atmospheric ambient art canvas when idle, and an explainable, instant consensus engine when everyone sits down to watch.

What it does

Aura Vega TV transforms Fire TV into a living room command hub and deterministic co-viewing consensus powerhouse:

  1. Circadian Ambient Hub: When idle, Aura transforms the television into an energy-conscious, atmospheric canvas with dynamic time-of-day circadian lighting (Dawn, Daylight, Sunset, Cosmic Midnight), real-time weather glance, and household member presence badges.
  2. Couch Consensus Engine: With a click of the Fire TV remote, viewers toggle active household members. Our deterministic multi-factor scoring engine aggregates group genre affinities, checks majority vetoes, factors in critical ratings (Rotten Tomatoes & IMDb), and applies real-world environmental context (time of day, weather, bedtime limits).
  3. Transparent Explainability ("Why This?"): No black-box AI opacity. Viewers see clear, plain-English explanations in under 5 seconds (e.g., "Unanimous match across all active viewers • Fits late-night 2-hour window • 94% Rotten Tomatoes").
  4. Instant Hardware Playback & Fire TV X-Ray: Press Select to immediately stream 4K HDR10 content with Dolby Atmos via Vega OS hardware-accelerated W3C media surfaces, complete with Fire TV X-Ray actor and soundtrack insights.

How we built it

  • Target Platform: Amazon Fire TV powered by the official Amazon Vega OS SDK 0.24 (Kepler platform).
  • Core Runtime: React Native 0.83 (Bridgeless Architecture) with dual application targets:
    • UI Target: com.amazon.kepler.runtime.react_native_kepler_4 (index.js)
    • Headless Service Target: com.amazon.kepler.runtime.react_native_kepler_4 (service.js under process group main_and_service)
  • Official Amazon Vega SDK Integrations:
    • @amazon-devices/react-native-w3cmedia: Hardware KeplerVideoSurfaceView and headless VideoPlayer controller.
    • @amazon-devices/vega-carousel: TV Carousel v2 with custom CarouselItemDataAdapter for 60fps horizontal shelf navigation.
    • @amazon-devices/kepler-cli-platform: Metro bundler toolchain and Static Hermes compiler.
    • @amazon-devices/kepler-a11y-settings-interface-turbo: TV accessibility and closed-caption styling TurboModule bridge.
  • Enterprise Microservices Cloud Stack: Node.js / TypeScript, Supabase PostgreSQL with pgvector HNSW semantic embeddings (768-D), Appwrite multi-profile authentication, AWS CloudFront edge streaming, and Kong API Gateway.
  • Performance & Testing: Compiled to Static Hermes bytecode v96 for sub-second cold starts. Verified with 22 automated Jest test suites and 166 unit tests covering 7 distinct conflict scenarios, 15 adversarial QA edge-case boundary tests, and 100% strict TypeScript type checking.

Challenges we ran into

  1. Path-with-Spaces Windows Metro Bug:
    • The @amazon-devices/kepler-compatibility-metro-config/dist/src/utils.js failed to execute Metro plugins when Windows user paths contain spaces (C:\Users\Ronak Jain\...). We diagnosed the root cause and patched the command invocation with proper quoting and node prefix.
  2. Windows Hermes Compiler Missing Binary:
    • The default @amazon-devices/kepler-cli-platform threw Unsupported host OS: win32 during Hermes compilation. We resolved this by linking hermesc.exe from hermes-compiler/hermesc/win64-bin, enabling native Windows bytecode generation.
  3. Carousel v2 Typing Incompatibilities:
    • Migrated from legacy FlatList to the official Vega Carousel v2, resolving TypeScript constraints by implementing CarouselItemDataAdapter<MediaItem, string> with strict key extraction.

Accomplishments that we're proud of

  • 100% Deterministic & Honest: Built a genuine multi-viewer consensus engine with zero fabricated metrics, zero hallucinated LLM wrappers, and mathematical explainability.
  • Production-Grade Release Pipeline: Clean Metro release bundling, Static Hermes bytecode compilation, official Vega SDK package generation, and zero-warning TypeScript validation.
  • Exhaustive Testing: 166 passing automated tests across 22 suites, including adversarial QA for majority vetoes, tie-breakers, and focus trap prevention.

What we learned

  • 10-Foot UI Ergonomics: Designing for television screens requires completely different rules than web or mobile: high-contrast 3px glowing focus rings (#FF9900), strict 80dp/60dp safe zones to prevent overscan clipping, and tactile 2D Cartesian D-pad spatial navigation.
  • Power of Vega OS: Static Hermes bytecode provides smartphone-fast cold boots on lightweight TV hardware, while decoupled hardware W3C media surfaces guarantee buttery 60fps UI responsiveness during 4K video playback.

What's next for Aura Vega TV — Living Room Consensus Engine

  1. Household Voice Profiles: Integrating Alexa voice identification on Fire TV remotes to recognize individual voices and automatically set active co-viewers when they speak.
  2. Streaming Catalog Deep-Linking: Direct intent URI deep-linking into Amazon Prime Video, Netflix, and Max apps on Vega OS.
  3. Watch History Sync: Real-time cross-device sync to ensure recently watched movies are automatically excluded from group recommendations.

Built With

Share this project:

Updates

Submission history