How I built a Plex server that recommends movies better than Netflix

We all rely on recommendation engines, whether they are suggested by streaming services or curated by media platforms, to guide our entertainment choices. Yet, for many users, this experience often falls short of being truly personalized.

The frustration lies in a fundamental gap: these systems frequently fail to capture the nuanced why behind a user’s enjoyment. They recommend titles based on simple metrics, but they miss the deep context—the specific emotional connection or thematic resonance that truly defines a great movie experience.

Existing recommendation engines often treat viewing history as mere data points rather than a complex tapestry of personal taste and preference. The result is algorithms that are helpful on a superficial level but ultimately miss the mark when suggesting something genuinely novel or satisfying.

This feeling led to a pivot. Tired of passive consumption guided by imperfect technology, the focus shifted from accepting what was offered to actively creating what was desired. The realization dawned that if the systems we use don’t understand our true cinematic desires, why not build a system that does?

This realization sparked an ambitious project: the decision to develop a custom recommendation engine powered by artificial intelligence. The goal was simple yet profound—to marry the vastness of personal viewing history with the power of AI to generate truly insightful and meaningful suggestions.

By leveraging personal Plex watch history, the aim was to move beyond algorithmic guesswork. Instead, the system would analyze not just what we watched, but how we watched it, identifying complex patterns that human curators often struggle to articulate.

It is a project born from dissatisfaction, driven by the desire for control and deeper understanding in the digital age. It represents the journey of turning passive consumption into active creation—a testament to the power of applying technology not just to filter content, but to understand taste itself.