All Activity

Fair enough on giving it a spin because it's open source — that's exactly how it should work. But "AI checking AI" isn't actually what happens in a serious workflow. The AI writes; the checking is done by things that can't hallucinate: the compiler, the type checker, the test suite, and a human who reads the diff before merging. That's the same pipeline human-written code goes through, because humans make plenty of mistakes too — that's literally why code review and CI exist in the first place. Comparing it to AI search making things up mixes two different problems. A made-up citation can slide past you because there's no automatic way to catch it. Made-up code gets caught the moment you run it. The safety net isn't trust in the AI — it's a process that assumes *every* author, human or not, gets things wrong.

I think it's important to mention that, in my case, the damage wasn't limited to incorrect artwork while EmbyIcons was installed. Even after uninstalling EmbyIcons and testing the "fixed" versions, the library remained permanently corrupted. Every time I updated the artwork for a movie, those new images would propagate to other unrelated movies. The corruption was persistent and couldn't be repaired with metadata refreshes. In the end, the only solution was to completely remove my entire media library (around 9,000 items) and import everything again to rebuild the database from scratch. So while the root cause may indeed have been Emby assigning 00000000 IDs, the end result was a permanently corrupted library that couldn't be recovered.

The fact that it's open source is the only reason I was willing to give it a spin. As for the analogy not holding up yes it does cause at this point with all the mistakes AI has made, all the mistakes AI devs have made, AI search making up what they don't know using AI to check AI's work is just a bad idea.

Yep, Claude Opus is an AI model, but the analogy doesn't really hold up. It's made by Anthropic, whose entire business *is* building AI models — it's not like an OS vendor bundling a half-baked antivirus as an afterthought. If anything, it's the opposite of your example: I'm going to a company whose one and only product is the tool I'm using. Also, using AI as a programmer isn't about blindly trusting it. It's a tool, same as a compiler, a linter, an IDE with autocomplete, or Stack Overflow back in the day. I still read, understand, and test everything before it ships. The skill hasn't disappeared — it's just shifted toward reviewing, architecting, and knowing when the output is wrong. Judging devs for using AI is like judging carpenters for using power tools. By the way, the project it's fully open source — the entire codebase is up on GitHub for anyone to inspect. If you have concerns about quality or security, the best way to back them up is to actually look at the code. If you find something questionable, point it out (or even open an issue) — that would genuinely help the project and I'm sure the maintainers would welcome it. That's the nice thing about open source: nobody has to speculate. The code speaks for itself.

I'm assuming Claude Opus is another AI? If so that's just like using the security tools Microsoft provides in Windows instead of going out and getting security apps that don't benefit them. From a security standpoint I've never thought it was a good or smart idea to use the tools created by a OS manufacture that has nothing but problems with their OS to make sure you're safe.

I am having problems connecting to server. I’m not sure if it’s because of my membership fees or what if it is how do I take care of that so I can get service again?

I can report some good news regarding the latest update (2.2.56): I now can play video files that are 1080p H264 with AAC audio. Sadly this is were the good news end. AV1 (be it 1080p or 4k) with OPUS audio still fails, as do MP3 and FLAC audio files. All of them lead to an immidiate crash. WAV does not crash, but result in the electronic "fingernails on a chalkboard" I mentioned in my first post in this thread. AAC audiobooks inside a m4b container also lead to a crash. So do HEVC video files. I hope this helps narrowing down the issue. MacOS version is 26.6 if that is relevant.

Hi Luke, I left the app open for a while and the movie eventually finished downloading. Is it expected to take such a long time to sync? I am attaching an updated log file because I can see a lot of "task failed" entries that don't seem expected. Additionally, as a workaround, I tried converting the movie to the desired quality, which left me with two versions (the original and the converted one). However, there appears to be no way to select and download the converted version specifically. Thanks! emby_android_1785776838860.txt

Jolly good. Similar for movies, like say Ben Hur. Brackets don't seem to be essential. I have Ben Hur (1959) Part 1 and Ben Hur (1959) Part 2 in the same folder and also get the additional part showing up.

No, not that I’m aware of. I never had problems watching movies so far. I did delete the automatic trigger for Scan media library, so the posters are still there. If I execute it manually, I always end with the same result, no matter the time of day.

Gotcha - I totally misunderstood what you were trying to say and I apologise.. I know I have the name correct because the first part is matching perfectly. I will try using brackets. --- Edit: Brilliant - that worked perfectly and, as you wrote (pun intended), the second file shows up as an "additional Part. I also did "The Plex dance" (removed it, rescanned, added back, rescanned) and it works for Plex as well. Thank you so much! ...for clarity incase someone else is interested, this was the format that worked (taken from @arrbee99's post): S01E01 name (Part 1).mp4 S01E01 name (Part 2).mp4

Yes, But I can not get the old name out of my head

Hokay. Just to mention I have the very first episode of Murder, She Wrote as given above (the second example, two S01E01's), and it shows up as Ep1 with an additional part.

Thanks, but it's a single episode in IMDB; I just have it across 2 files. Emby seems to think they are variations of the same episode rather than 2 parts of the same episode.

https://post.smzdm.com/p/apvok92x/?sort_tab=hot/#comments

Might depend on how its named by TVDB or whatever ? Maybe - S01E01 name (Part 1).mp4 S01E02 name (Part 2).mp4 or S01E01 name (Part 1).mp4 S01E01 name (Part 2).mp4 Personally, if testing I'd remove item from Emby, re-scan Emby db, delete all info except the item itself, re-add and re-scan Emby db.

agh, OK. No, I had already tried that before posting and it didn't work. I tried " - Pt1" (Plex style) and " Pt 1" at the end. Thanks anyway.

Hello Gyvari, ** This is an auto reply ** Please wait for someone from staff support or our members to reply to you. It's recommended to provide more info, as it explain in this thread: Thank you. Emby Team

i just realized that on tv shows library, director and crews not picking up tmdb id, is it intended like that? i have to identify manually each one and it's a lot of work. movies library is good though, all actor and crews tmdb id is always filled.

Is there a way I can change the network settings on Emby Web to achieve this? Thanks

Local: port 8098. Remote: port 8098. Both show 8098 — not the 8096 I manually set. Luke Has confirmed my suspicions. Thanks

0.0.012 i noticed a library i had renamed became 'stuck' in the config Now keys enabled libraries by id (rather than name) in the config.xml MIgration pathway to new ID based config 2026-08-05 04:41:55.736 Info Poster To Folder: Poster To Folder: migrated 25 legacy library entries to LibraryId (24 matched by name to a current library, 1 orphaned - no current library named the same, left untouched)

Still is not working. I've updated iOS and even tried the iOS 27 beta. Tailscale is updated as well.

Thanks Luke. So essentially, if I don't allow any files within D:\Root itself? So adjusting my above example, Library A would have: D:/Root/OldRoot D:/Root/Folder1 D:/Root/Folder3 And Library B would have: D:/Root/Folder2 This way there's no overlap?

Hi, what is "this" exactly that you're asking about?