Known good hardware stacks

[jritsema]

I have a custom AI Agent HTTP API that I want to put a voice interface on top of. OVOS looks like a good open source software stack to build this on. I've seen the supported devices page, but it would be great if there was a list of known hardware stacks that people have proved work well (I haven't found many examples online). For example, a Raspberry Pi 4, 8Gb with a particular microphone and speaker configuration. Is there something like this out there?

[goldyfruit]

Most of the USB microphone will work as long as they are supported by the Linux kernel.

[jritsema]

Thanks for the quick response. That's helpful to know about USB microphone compatibility.

I completely understand that most Linux-compatible USB mics should work. What I'm hoping to find are some real-world examples of complete setups that people have successfully built - something like "Raspberry Pi 4 + USB Mic Model X + Speaker Y worked great for me." and ideally a video demonstrating the end experience.

As someone new to voice AI hardware, having a few proven configurations would give me confidence in my hardware choices and help avoid any unexpected compatibility issues during setup. I found the SmartGic YouTube channel which has some demos, but I'm wondering if there are other community resources where people share their working builds?

No worries if this doesn't exist yet. I know documentation takes time to build up in open source projects. Just thought I'd ask in case there's a community forum or wiki I missed!

[goldyfruit]

I don't think we have a page for that not a bad idea but might not be easy to maintain as we will not be able to test the hardware at each OVOS/Linux/Kernel releases.

FYI; Smart'Gic is my YouTube channel.

[Bulwagga]

I've been an OVOS/Mycroft hobbyist for several years now. I use a Raspberry Pi 4B, 8G of ram, with a Plantronics P3200 Poly USB speaker/mic. From a system point of view, this hardware has been very reliable. I have also run successfully for a while on a Pi 3B with a $3 plastic mic and a single wired earpiece. Occasionally I have had to play with alsamixer to assign the devices correctly, but no real problems. Early on I was running on a Dell PC and an HP laptop both running Ubuntu. The complications usually come from the software stack side. Because the software side is a long chain of sequential processes that are rapidly evolving, you need to have appropriate backups in mind for each step. Depending on your tolerance for sending your communications to multi-nationals for speech-to-text and text-to-speech translation, the docs are a good guide to setting up a reliable software stack.