LABORATORY
GREY

Evergreen Media Server

A custom web interface for streaming video from a local Raspberry Pi server.

last updated: 2024/09/19

During the pandemic in August of 2020, I decided to build a local media server. Rather than using an off-the-shelf NAS solution, I set up a simple Raspberry Pi web server that hosts my media (and this website!).

I had several objectives:

• Create a web, mobile, and TV-friendly CSS layout.
• Implement a process to legitimately purchase and transcode video content (1080p, 540p, 288p).
• Develop a custom metadata template for filtering content.
• Support time-based seasonal content.
• Streamline adding and expanding content.
• Ensure TV view supports remote control.
• Devise a backup strategy.

I wanted to keep it simple but also incorporate pet features I'd been wanting. Managing the metadata personally has been fun; I choose not to import any online data, logging details myself about each show. This also enables some unique dynamic playlists.

The seasonal content approach is a great way to omit episodes you only want to see around a holiday, with some playful theming for each playlist. For instance, clicking the pumpkin emoji (seen in the screenshot) brings you to a Halloween-themed playlist with an orange color scheme, showing all episodes tagged for "Halloween."

The server setup itself is a separate topic, but to note one detail: remote support is provided courtesy of Flirc and an Apple TV remote mapped to keystrokes, with JavaScript handling movement between UI sections.

Conclusions

The server has been running smoothly. I consider this project a success-I use it almost daily. It's exactly what I want it to be, and it works without issue. I still plan to add more advanced features like complex searches or a biography feature with creative credits.

The code is straightforward enough for me to maintain and adapt. I don't typically love web development, but with clear constraints and goals, it can be enjoyable. It's actually how I first learned to program (making games in QBasic for fun).

The process for adding and expanding content is ongoing. I've developed a series of Python scripts to handle data integration and ensure files conform to naming conventions. This is likely the biggest remaining pain point, but it has improved a lot over time.

I haven't made this code public, but if I do, I'll add a link here. I also intend to document more details of the server's setup on this site eventually.