A Netflix-grade streaming platform for Apple TV and iPhone — billboard heroes, focus-driven rows, profiles and Continue Watching — rebuilt screen by screen to match the experience people already know.

Overview
StreamNest is a full streaming service — a native Apple TV app, an iPhone app, a TypeScript API and a web admin console — built to feel exactly like the big-name streaming apps viewers use every day.
On the TV, viewers pick a profile and land on an inset billboard with title art, ratings and a synopsis, then browse genre rows where the focused tile expands into a wide still with details underneath. Dedicated Shows, Movies and My StreamNest tabs, on-screen keyboard search, Continue Watching with progress, My List, and a full-screen player round out the experience. The iPhone app adds downloads for offline viewing, and an admin console manages the library, users and background jobs.
Key Features
Challenge
"Make it look like Netflix" is easy to say and hard to deliver on a TV. The 10-foot interface lives or dies on focus behavior — how tiles grow, how rows snap, what the remote does at every edge — and the details that make it feel premium are rarely documented. On top of the UI, the platform needed a catalog with real artwork and metadata, smooth seeking over the internet, multiple profiles, and a backend that could scale media processing without a big cloud bill.
Solution
I treated the reference app as a spec. An automated XCUITest walk captured every screen of the reference experience on a real Apple TV 4K along with its accessibility tree, giving exact measurements — 68 pt margins, 310×504 tiles that expand to 900×504 on focus, 24 pt corner radii, a 112 pt nav bar. Every gap between that and StreamNest was logged in a parity matrix and closed in SwiftUI: a custom top navigation bar replacing the system tab bar, row-snapping vertical scroll, an info panel under the focused row, inset billboards with title logos, and a Netflix-style profile picker. The backend is TypeScript on Bun and Fastify with SQLite via Drizzle, TMDB metadata and artwork, HTTP range streaming for instant seeking, and a typed API contract shared across every client. Heavy media processing fans out to a fleet of Raspberry Pi workers over Cloudflare Tunnels, each adding capacity with a one-command installer.
Results
StreamNest runs in production on Apple TV and iPhone and is close to indistinguishable from the app it was modeled on — the same layout, motion and remote behavior, under its own brand. The measured parity process is now a repeatable playbook: given any app or website, I can capture it, measure it, and rebuild it faithfully for a client's own product and content. Adding a Raspberry Pi to the worker fleet adds processing capacity in minutes, keeping running costs to a home server and a few small boards.
Gallery






