Automated Testing Layers. Fast-tracking quality for hundreds applications

A case study explores the Streaming company's journey of building automated testing layers within a complex, multidimensional React monorepo and achieving quality contribution feedback within minutes.

Zattoo is a streaming service that develops multiple projects that serve almost all possible technologies that could play video. It is customizable for dozens of white-labeled customers and scales to hundreds of unique end-user applications. With multiple teams contributing daily and regular release cycles, our biggest challenge is ensuring fast, reliable, and maintainable quality assurance across the entire codebase.

This talk details how we transformed our approach to testing, moving from a labor-intensive, manual acceptance process to a fully automated quality feedback loop that provides results in less than three minutes for every contribution.

The Fourth Platform: How Vega got us surprisingly close to “Write Once, Run Everywhere”.

“Write once, run everywhere” is a promising goal in software development, but one that often breaks down under real-world problems. At Zattoo, building streaming applications across Android, Apple, and Web meant years of separate native stacks that were not aligned, and therefore did not scale. This talk explains how moving to a multiplatform architecture, enabled by Vega as a fourth (React Native first platform), brought us very close to achieving that goal.

In this talk, we will walk through the architectural decisions behind Zattoo’s transition from platform-specific native applications to a unified React Native based multiplatform codebase. I’ll explain how our codebase is structured, what parts of the application are shared, where platform-specific boundaries still exist, and how Vega fits into this architecture alongside Android, Apple, and Web.

We will also discuss the practical trade-offs of this approach, including challenges encountered during the migration, limitations of cross-platform abstractions, and areas where platform-specific work remains necessary.

The session will also cover how teams are organized around business domains rather than platforms, and how this shift helped improve alignment and speed up delivery across ecosystems.

Rather than presenting an idealized solution, this talk focused on lessons learned from building and shipping a real production streaming application across all major ecosystems, and what it takes to make a “write once, run everywhere” strategy look in practice.

This talk is intended for frontend and platform engineers working on applications that must run across multiple ecosystems, including Mobile, Web, and TV devices. It will be particularly relevant for teams evaluating or already using React Native, as well as for engineers interested in unifying business logic, UI, and tooling across platforms without sacrificing performance or platform integration.