User builds 15MB all-in-one media server app using Claude

Consolidating media tools

For people who run their own home media servers, managing content can become a technical headache. Typically, users have to run a suite of different applications like Sonarr, Radarr, Lidarr, bazarr, Prowlarr, and Seerr to handle various tasks. Each of these separate tools has its own unique user interface and operating system, which adds layers of complexity to the setup.

These individual apps can be quite heavy, often taking up between 110mb and 140mb of space each. This fragmentation means that a user's server can become cluttered with multiple running processes just to organize a movie or TV show library. It creates a fragmented experience where the user must jump between different windows and menus to manage their media.

Reddit user CoderLuii sought to solve this problem by creating Holyarrs. As reported by XDA, the goal was to combine the functionality of all those disparate tools into one single web app. The developer wanted to create a unified experience that feels more like a consumer streaming service than a collection of backend management tools.

The vibe-coding process

The creator built Holyarrs using a method described as 'vibe-coding' with Claude. This approach involves using an AI model to generate code based on high-level descriptions and goals rather than writing every line of logic manually. To ensure the final product was efficient, CoderLuii placed very strict limits on the app's memory usage, boot speed, and overall file size during the development phase.

The developer did not just take the first version the AI produced. They actually asked Claude to build the Plex-like app in two different ways to compare the results. One of these versions was found to be 3x lighter and faster than the other, leading the developer to choose that specific path for the final product.

This iterative process with the AI allowed for a high degree of optimization. By setting hard constraints on performance from the start, the developer ensured the app wouldn't become bloated. The result is a tool that manages 12,800 titles while remaining extremely lean in terms of system resources.

A Spotify-inspired UI

One of the most distinct features of Holyarrs is its visual design. The developer intentionally modeled the user interface after Spotify, the popular music streaming platform. This choice was made to provide a familiar and polished look for users who are used to modern streaming layouts but want the power of a self-hosted server.

Because it is a web app, it avoids the need for a dedicated operating system for each tool. It functions as a single point of entry for the user, mimicking the feel of a service like Netflix. This design philosophy aims to hide the complexity of the backend tools while keeping the core features accessible in one place.

The community response to this aesthetic and functional blend has been very positive. Users on Reddit have noted that the project stands out because it is a functional, well-thought-out application rather than a simple prototype. The developer has also shared a full roadmap on the official website to show the project's long-term direction.

Technical specs and future

The technical efficiency of Holyarrs is one of its most impressive traits. Despite its ability to index a massive library of 12,800 movies and TV shows, the app only requires 70mb of RAM. This is significantly lower than the resources required by the individual apps it replaces, making it ideal for users with limited hardware.

While the app is currently available for use, it is not yet open-source. However, the developer has stated that the project will be open-sourced in the future. This transparency suggests a commitment to the community and a desire for the project to grow with outside input and peer review.

Moving forward, the developer has big plans for the app's expansion. XDA will be following the project closely as it evolves. For now, it serves as a notable example of how AI can be used to simplify complex workflows into highly optimized, user-friendly software.