iOS Factory Pipeline

The iOS Factory Pipeline is the central artifact of the Kelly Factory — a structured 10-step operating system that processes opportunities through research, design, build, test, and release with quality gates at each transition. It is the operational definition of the factory: what gets built, in what order, with what validation, and how quality is enforced at each stage.

The pipeline consists of these steps:

Step 0: CHECK STATUS — Verify the factory state, confirm pipeline readiness, confirm all prerequisites are in place before beginning a new build cycle. No work begins until status is confirmed clean.

Step 1: DISCOVER — Research Station runs opportunity scan across Reddit and App Store. Identifies potential niches, scores them, flags those worth pursuing. This is the factory's input function — the queue of work to be done.

Step 2: VALIDATE — Discovery Gate. The opportunity must score 28 out of 40 on the validation rubric before proceeding. This gate ensures no low-quality opportunities consume factory resources. (See discovery-gate)

Step 3: DESIGN — Design Station generates UI specifications from the validated opportunity. Produces mockups, design system, asset specifications. Must score 95 out of 110 before BUILD begins. (See design-gate)

Step 4: BUILD — The build phase where code is written. Multiple sub-agents work in parallel (see sub-agent-parallelism) to produce the complete app. SwiftUI, SwiftData, StoreKit 2, dark mode, privacy policies — all produced end-to-end by the build team.

Step 5: POLISH — Post-build refinement phase. Haptics, confetti, spring animations, visual polish, and performance optimization. Must clear the 70% quality threshold before SUBMIT. (See polish-gate)

Step 6: LAUNCH — Final pre-submission validation. App Store metadata finalized, assets uploaded, signing configured. Readiness check before the formal submission.

Step 7: SUBMIT — Automated submission to App Store Connect. The app goes live for Apple's review process.

Step 8: MARKET — Marketing Factory activates. Twitter thread published, TikTok scripts deployed, Instagram carousels posted, landing page goes live. The product's marketing starts the moment it's submitted.

Step 9: TRACK — Performance monitoring. App Store rankings, download rates, review sentiment, revenue. Data flows back to inform future Discovery phases.

Step 10: LEARN — Post-review retrospectives. Captures wins and failures from this build cycle, feeds them into learning-loop so the factory gets smarter for the next iteration.

Quality gates are the structural backbone of the pipeline. They are enforcement mechanisms that AI cannot bypass — defined criteria checked by the Router (not the builder) before the pipeline advances. The builder cannot pass its own gate. This structural separation between execution and validation is what makes the quality system trustworthy rather than self-serving.

The pipeline is designed for parallel execution at scale. When 20 apps are in the pipeline simultaneously, each is at a different step — some in Discovery, some in Build, some in Marketing. The Router tracks state across all active products and routes attention to whatever needs it. This is what enables the "24 products in 24 hours" output — not mockups or landing pages, but real production-ready products with full-stack apps, AI integrations, payment systems, and live demos.