Learn and Be Curious: App Platform (iOS App Factory)

A shared backend, SwiftUI app template, and CI/CD automation that turns per-brand manifests into shipped iOS apps.

Overview

Learn and Be Curious is a platform-level iOS app factory designed to eliminate the overhead of repeatedly building the same foundations for each new app. Instead of cloning and modifying codebases, the platform centralizes everything that should be consistent across apps and allows each brand to vary behavior through configuration.

The platform is composed of three core parts:

Platform backend: A multi-tenant AWS backend providing shared authentication, APIs, data storage, and operational defaults, versioned and deployed via CI/CD.
SwiftUI template app: A production-ready iOS starter with a manifest-driven configuration system, navigation shell, auth/session handling, and a typed API client.
Per-brand app manifests: Lightweight brand repos that define identity, theming, versions, legal links, and feature flags—with optional overlay code for app-specific behavior.

Core capabilities such as authentication are always enabled and treated as platform guarantees, while optional capabilities are controlled through feature flags. This allows each app to selectively enable functionality—such as feedback flows, push notifications, deep linking, sharing, image capture, ratings prompts, or AI features—without branching the underlying codebase.

At build time, CI applies each brand’s manifest to the shared template—configuring bundle IDs, assets, Info.plist values, versions, and feature flags—to produce a fully signed, App Store–ready iOS application.

Built with

Idea: ChatGPT was used to brainstorm and evaluate approaches for shipping many iOS apps without re-implementing shared functionality, leading to a manifest-driven app factory design.
Graphic Design: The VS Code + OpenAI Codex extension (GPT-5) was used to generate and refine theming and asset integration patterns, enabling branding through manifests rather than manual design tooling.
Architecture Design: The VS Code + OpenAI Codex extension (GPT-5) was used to design and implement the full platform architecture, spanning the shared backend, SwiftUI template app, per-brand repositories, and build-time composition flow.
Requirements: The VS Code + OpenAI Codex extension (GPT-5) was used to translate high-level goals into concrete, testable requirements, including feature flags, manifest schema, authentication flows, and a repeatable new-app launch process.
Development: The VS Code + OpenAI Codex extension (GPT-5) was used to write and iterate on all production code across the system, including iOS SwiftUI code, backend services, build scripts, and GitHub Actions CI/CD workflows.
Testing: The VS Code + OpenAI Codex extension (GPT-5) was used to write unit tests, execute test suites, interpret failures, and refine code until builds and tests passed end to end.
Deployment: The VS Code + OpenAI Codex extension (GPT-5) was used to implement and validate deployment pipelines, including AWS CDK for backend infrastructure, GitHub Actions workflows for CI, and signed iOS builds for TestFlight and App Store distribution.

AI inside

The iOS template includes a feature-flagged AI Playground screen used to validate AI integration at the platform level. When available, it exercises Apple’s on-device Foundation Models (SystemLanguageModel via FoundationModels). On simulators or unsupported devices, the platform falls back to deterministic local behavior so AI-dependent code paths remain testable and predictable.

Tech Stack

iOS app template: Swift, SwiftUI, XCTest
Authentication (mobile): Amazon Cognito Hosted UI (Sign in with Apple + Google), custom URL scheme redirects
Backend IaC: AWS CDK (TypeScript)
Backend services: Amazon API Gateway (HTTP API), AWS Lambda, Amazon DynamoDB
Edge & static assets: Amazon S3, Amazon CloudFront
Observability: Amazon CloudWatch logs and alarms
CI/CD: GitHub Actions (reusable workflows, xcodebuild test/archive, artifact uploads)
Build tooling: Bash scripts, jq (manifest validation), rsync (scratch copies), PlistBuddy and perl (project and Info.plist rewrites)

Next Steps

Extend the app factory to Android by creating an Android template app and manifest-driven branding pipeline (theme, feature flags, bundle/application IDs, assets), plus matching CI/CD workflows to ship multiple branded Android apps with the same platform approach.

← Back to all projects