Preferences
Theme
Previews play by default. Turn this off to pause them; reduced-motion preferences are always respected.
← Projects

03 / Portable media pipeline

YarReader

A portable media library that works without a server.

Primary capabilityOffline, recoverable media pipeline

What it is

A complete working system.

A browser-based reading experience backed by a crash-recoverable pipeline that converts mixed publication formats into a verified, self-contained offline library.

Why I built it

The engineering question.

Portable archives fail when readers depend on a database, a network, or fragile application state. YarReader pushes complexity into the build pipeline so the activated library stays ordinary, durable, and movable.

Engineering highlights

What the project demonstrates.

  • CBZ, CBR, EPUB, PDF, and image adapters
  • Content-addressed normalization
  • Crash-recoverable transactions
  • Immutable static exports with an offline reader
Architecture / technology

How the project is structured.

Primary capability
Offline, recoverable media pipeline
Technologies
  • TypeScript
  • CLI
  • Static HTML
Characteristics
  • Content addressing
  • Crash recovery
  • Verified exports
  • Offline-first
InspectOpen each source safely and record what it contains
IdentifyMatch the book using rules first, optional AI second, and a person when needed
ConvertTurn pages into consistent images with digital fingerprints
PublishCheck the entire new library before making it the active copy

Go deeper

Overview first. Evidence when you want it.

The case study explains the architecture, boundaries, tradeoffs, and current state. The running application and repository provide the proof.