Bluebell SDK

Kotlin Multiplatform foundation for scalable app teams

Bluebell is an opinionated SDK for teams that want consistency without friction: lightweight configuration, architecture plumbing that stays out of your way, Compose UI-oriented development across platforms, and extension points ready for AI-powered workflows.

View Bluebell on GitHub Open Eyespie demo app
Kotlin Multiplatform Compose UI Flux-style state Configuration plugin AI-ready

Reference Demo: Eyespie

KMP app · Android + iOS · AI-assisted gameplay

Eyespie is the first public demo app using Bluebell. It is a Kotlin Multiplatform "I spy" game that uses shared Compose UI, camera and location capabilities, Supabase-backed services, and MediaPipe-based vision / GenAI hooks to explore offline-capable, AI-assisted mobile application patterns.

View demo source Read demo overview

Project Charter

Charter v1.0 · Effective March 4, 2026
  • Configuration is plugin + files + environment, by convention first.
  • Architecture and design patterns are standardized, not optional.
  • UI stays Compose-first across shared and platform targets.
  • Data pipelines must be reliable from domain to presentation.
  • Security defaults are mandatory, including store attestation leverage.
  • Observability and structured logging are required by default.
  • Extension points must support AI-enabled capabilities over time.

UI State Flow

Unidirectional immutable loop for state:

Actions Reducer Store UI

External Data Flow

Bi-directional immutable path for externals to the app:

device / api / database / cache data source repository use cases user

Unobtrusive Configuration Plugin

  • Use convention-based defaults through the build plugin first.
  • Layer configuration files and environment values cleanly.
  • Generate typed runtime configuration for application code.
  • Support local overrides without rewriting core behavior.

Architecture & Design Patterns

  • Provide clean boundaries between app layers.
  • Keep UI patterns Compose-first across shared and platform code.
  • Build reliable data pipelines from domain to presentation.
  • Offer reusable building blocks instead of hard lock-in.

Extensible, AI-Enabled Core

  • Expose extension seams for new features and integrations.
  • Support vision and GenAI integrations in app modules.
  • Design for evolving capabilities without framework churn.