A suggestion for Berry Audio: Keep it lightweight and make it agent-friendly
Posted: Sat Oct 10, 2026 5:09 pm
Hi all,
First of all, I've been a long-time Volumio and moOde user, and Berry Audio has been a breath of fresh air. For some reason, I'm hearing a clear improvement in sound quality, and I'm really enjoying using it.
I'm also a massive Radio Paradise listener. One of the reasons I used Volumio over moOde was its Radio Paradise plugin and API, which gave me access to lossless streaming rather than lossy streams.
That brings me to my suggestion for Berry Audio.
Why not keep Berry Audio as bare-bones and lightweight as possible, but make it easy for AI agents to extend and customize?
If the platform had well-maintained, accessible Markdown documentation files explaining its architecture, configuration, interfaces and extension points, an AI coding agent could use those files to help users build their own features and fix their own problems.
I built a Radio Paradise extension for Berry Audio in around 20 minutes. It probably would have been even quicker if the right Markdown documentation had been available for my agent to work from.
I think this approach could open up some really interesting possibilities.
For example, I don't need every source Berry Audio offers. I mainly need Library, Storage, Radio and Settings. Rather than having to scroll through options I'll never use, I can ask my agent to remove the ones I don't need.
Imagine if Berry Audio had an agent skill or extension that walked users through setting up their own installation. It could ask questions such as:
- Will you be using DSP? No? Fine, let's remove or disable it.
- Will you be using Bluetooth? No? Let's leave it out.
- Do you need these additional sources? If not, let's simplify the interface.
The same principle could apply to adding functionality. When I uploaded my storage, I couldn't find a way to synchronise my library. There doesn't appear to be a library sync button where I need one. An agent could investigate how the library works and help build that missing functionality, rather than requiring me to wait for a developer to implement it.
Obviously, there would need to be sensible limits. Some features are fundamental to the platform, and not everything can safely be removed or changed. But with the right documentation and clearly defined extension points, users could customise their installations without the core project becoming bloated.
My suggestion is simple: keep Berry Audio small, fast and focused on doing one thing exceptionally well — playing music. Let the community and their AI agents take care of the extras.
Rather than trying to build a platform that does everything for everyone, give users a solid foundation and the tools to make it do exactly what they want.
I suspect this could become an increasingly important way for people to install, extend and maintain software, and Berry Audio seems like an ideal project for that approach.
Just my thoughts, but I'm really enjoying the platform so far. Keep up the great work!
Cheers
Matt
First of all, I've been a long-time Volumio and moOde user, and Berry Audio has been a breath of fresh air. For some reason, I'm hearing a clear improvement in sound quality, and I'm really enjoying using it.
I'm also a massive Radio Paradise listener. One of the reasons I used Volumio over moOde was its Radio Paradise plugin and API, which gave me access to lossless streaming rather than lossy streams.
That brings me to my suggestion for Berry Audio.
Why not keep Berry Audio as bare-bones and lightweight as possible, but make it easy for AI agents to extend and customize?
If the platform had well-maintained, accessible Markdown documentation files explaining its architecture, configuration, interfaces and extension points, an AI coding agent could use those files to help users build their own features and fix their own problems.
I built a Radio Paradise extension for Berry Audio in around 20 minutes. It probably would have been even quicker if the right Markdown documentation had been available for my agent to work from.
I think this approach could open up some really interesting possibilities.
For example, I don't need every source Berry Audio offers. I mainly need Library, Storage, Radio and Settings. Rather than having to scroll through options I'll never use, I can ask my agent to remove the ones I don't need.
Imagine if Berry Audio had an agent skill or extension that walked users through setting up their own installation. It could ask questions such as:
- Will you be using DSP? No? Fine, let's remove or disable it.
- Will you be using Bluetooth? No? Let's leave it out.
- Do you need these additional sources? If not, let's simplify the interface.
The same principle could apply to adding functionality. When I uploaded my storage, I couldn't find a way to synchronise my library. There doesn't appear to be a library sync button where I need one. An agent could investigate how the library works and help build that missing functionality, rather than requiring me to wait for a developer to implement it.
Obviously, there would need to be sensible limits. Some features are fundamental to the platform, and not everything can safely be removed or changed. But with the right documentation and clearly defined extension points, users could customise their installations without the core project becoming bloated.
My suggestion is simple: keep Berry Audio small, fast and focused on doing one thing exceptionally well — playing music. Let the community and their AI agents take care of the extras.
Rather than trying to build a platform that does everything for everyone, give users a solid foundation and the tools to make it do exactly what they want.
I suspect this could become an increasingly important way for people to install, extend and maintain software, and Berry Audio seems like an ideal project for that approach.
Just my thoughts, but I'm really enjoying the platform so far. Keep up the great work!
Cheers
Matt
