Beta Testers Wanted

[jvgomg]

Calling All iPod Enthusiasts: Beta Testers Wanted

podkit is a brand new project and I'm looking for a small group of people who'd be willing to help test it before it's ready for a wider release.

What I'm looking for

You don't need to be a developer. You just need:

• An iPod you're willing to wipe — things might break, databases might get corrupted, and you may need to restore your iPod during testing. Don't use an iPod with a library you can't afford to lose.
• A music collection — local files or a Subsonic-compatible server (Navidrome, etc.)
• Willingness to give feedback — tell me what works, what doesn't, what's confusing, what's missing. Honest, detailed feedback is the most valuable thing you can contribute.
• macOS or Linux — Windows isn't supported yet

What you'd be testing

• Installing podkit and its dependencies
• Setting up collections and devices
• Syncing music (and video, if you have a video-capable iPod)
• The overall experience — is the CLI intuitive? Are error messages helpful? Does the documentation make sense?

What you'd get

• Early access to podkit before the public release
• Direct input into how the tool works — your feedback will directly shape the final product
• My genuine gratitude. Seriously, early testers are the people who make a project like this actually work.

Interested?

Drop a comment below and tell me a bit about your setup:

• What iPod do you have? (model, storage, any mods like iFlash?)
• Where do you keep your music? (local files, Navidrome, etc.)
• What OS are you on?
• What are you currently using to sync? (iTunes, Finder, Rockbox, something else?)

Even if you're not sure you have time, I'd love to hear from you. There's no commitment — just enthusiasm.

[DuncanGJ]

Hey just found this, trying it out later today. You have the exact same vision as me. Currently I have my library stored locally on external drives and then I am doing all of my library management with beets. Long term my vision is to have my local media server store all my music, have beets run a cron job nightly and be able to just plug my ipod into the server and have it sync. The same library would also be used by an instance of LMS for playing music at home.

Right now I have an iPod classic 5.5 gen with modded storage. IDK what the exact mods are because I bought it on ebay. I am close to hitting the storage limit, and the battery is of dubious quality so I will probably be doing some more modding soon. I also want to switch to a translucent faceplate for the y2k nostalgia bait.

I'm on Tahoe 26.3.1 (a), current sync process is pulling out my old windows tablet and syncing via media monkey (which isn't too bad a of an interface all things considered). iPod is apple software, and I plan to keep it that way.

[DuncanGJ]

Hey @jvgomg - of course! super stoked to contribute to this project in any way possible.

in terms of the docker question. Most likely, I still have not set up my server yet, and my priority right now is offloading movies from my seedbox for local playback and freeing up space.

I can report that I was semi-successful in getting things going last night. I created a few issues. But syncing was successful. The issues are as follows:

• the biggest issue was the toml config file is being read very inconsistently. If you pass a malformed config it will crash correctly, but values from properly formatted configs don't seem to be used. This meant that I had to use environment variables and specify things like my device with the -d flag during syncing.
• The next biggest issue was that my device (ipod 5.5, don't know the exact model num) was only being recognized as a generic iPod. I'm assuming that this caused podkit to fallback to the most conservative transcode settings, which meant that all of my ALAC files got transcoded into mp3's even though my device supports ALAC playback. Obviously syncing took a lot longer than it otherwise would've.
• I have a few minor issues that I'm going to add, but my iPod used to have space in its name, and that is not supported by podkit. The sync time estimate was way off (about 40 min for 4,500 songs that just needed to be copied, 75min for transcoding 1500 songs vs the 5hr45min estimated time).
• I need to investigate the cause of this further, but upon sync artists tagged as snoop dogg ft. dr. dre ended up as their own artist instead of under snoop dogg, which was the expected behavior. But that could also stem from how my files are tagged.

[DuncanGJ]

btw @jvgomg i really appreciated the --dry-run flag, super super useful, also keep rocking Sequoia , Tahoe sucks

[jvgomg]

Fantastic - thank you @DuncanGJ! Thank you for making issues. Good thinking. I'll follow up over there and we can get to the bottom of them.

I think your iPod not being sensed as a known model will likely be a product of iFlash models not being created with correct system data when initialised - the Apple software sees that the filesystem blocks are not the same as the original hard drive and doesn't write the system files like it normally would. This is the most important thing to work out. Either make the sensing profile the USB Device more or add some more magic to the podkit doctor command to repair/upgrade the SysInfo.

I imagine that the wildly incorrect time estimates are bad maths based on CPU calculations. I had to tweak it to be more accurate on my M2 Pro. What CPU are you using?

[DuncanGJ]

M5 pro,

I was wondering if it was due to issues with the iFlash. I bought my iPod off ebay with mods already done, so I'm not positive on what exactly was done.

[jvgomg]

Tahoe is the worst feature of an M5 😂. This aligns with your machine being faster and so over estimating, but 6 hours vs 2 hours is insane. Hopefully this is easy to fix. The logic is using the USB 2 bus transfer speeds and file sizes as the primary factor but does consider the CPU. I presume the end storage estimates are also way off and part of the picture.

[kylemcd]

hey there! so I was throwing my AI a problem i wanted to solve so that i could directly sync jellyfin to my ipod. it ended up using your library and i think a lot of what the agent wrote could probably be pulled into here directly instead (granted with a lot of cleanup since my repo is mostly slop). the major issue i kept running into was the sync process stopping mid way from what i assume was some kind of network issue with subsonic / jellyfin, not entirely sure there but it would just hang. for additional context my setup is music lib hosted on a NAS -> served through jellyfin -> subsonic plugin -> my macbook pro m1 pro -> podkit / my lib -> ipod classic 6th gen.

for context i just let my agent run with this and most of the layer it's written on top of podkit seems mostly redundant but it ended up working for what i was trying to do so yay i guess?

here's the main things the "jellypod" layer added:

• It makes Jellyfin/Subfin-to-iPod sync safer by default (device trust gate + safer first sync policy).
• It makes syncs more resilient (retry/backoff, stall detection/recovery, convergence re-passes).
• It adds persistent orchestration state around podkit runs (run history, trust state, first-sync state) using sqlite.

here's the code if you want to peruse https://github.com/kylemcd/jellypod