Best path for cleaning up duplicate People profiles / folders?

[Smitty018210 159]

I am working on a small helper app for my Emby People metadata.

The app scans the Emby metadata/people area and looks for possible duplicate person profiles. The goal is not to blindly delete anything. The goal is to help review possible duplicate People folders before doing anything risky.

Right now the app is looking for duplicates by things like:

• Same normalized person name

• Same TMDB person ID

• Same IMDb person ID

• Linked duplicate chains, where one profile matches by name and another matches by provider ID

• Skipped/problem profiles with missing or unreadable info

The issue I am trying to solve is duplicate People folders like this:

Vin Diesel
Vin Diesel-tmdb-12835

or cases where one profile has provider IDs and another is name-only.

My concern is this:

If I delete the “bad” / name-only People folder, will Emby just recreate it later from movie/show/episode cast metadata?

From what I understand, the person profile NFO itself does not store credits/media links. The media item metadata is what contains cast/actor information. So deleting a People folder may only delete the stored person metadata/image folder, but it may not fix whatever source caused Emby to create that person entry in the first place.

I am trying to find the safest and most Emby-correct cleanup path.

Questions for the Emby dev team:

• Is the correct preferred People folder format still like this?

Person Name-tmdb-123456
Person Name-imdb-nm1234567

• Should TMDB ID be treated as the preferred primary person ID, with IMDb as fallback?

• If a duplicate name-only People folder exists beside a provider-ID folder, what is the safest way to remove or merge that duplicate?

• Does Emby recreate name-only People folders when media NFO/cast metadata does not include a provider ID?

• If local NFOs are used, should the fix be made in the movie/show/episode NFO actor entries first, before deleting any duplicate People folder?

• Is there any safe way through Emby API or metadata refresh to tell Emby that one person profile should be treated as the correct one?

• Is there a recommended process for cleaning duplicate People records without damaging existing images, locked profiles, or manually edited person metadata?

What I am trying to avoid is building a delete/cleanup feature that appears to work, but then Emby recreates the same duplicate People profile on the next scan because the source metadata still points to a name-only actor entry.

Any guidance on the correct order of operations would help.

For example, should the safe process be something like:

• Identify the good provider-ID People profile.

• Check movie/show/episode cast metadata for name-only actor entries.

• Correct actor/provider IDs in the media NFOs first.

• Refresh/rescan Emby.

• Only delete the duplicate People folder after Emby stops recreating or using it.

I would appreciate any clarification on how Emby decides when to create a new People folder versus reuse an existing People profile, especially when one profile is name-only and another has TMDB/IMDb IDs.

[ginjaninja 609]

Is the end goal to ensure your library metadata is right/deduped or the metadata folder on filesystem? I would be tempted to focus on getting the library metadata correct and trust that emby will sort out the filesystem (apart from wasted space, does it matter?). That wont clean duplicates on the filesystem but at least your library will be 'correct'. I dont recall the api exposes the filesystem related metadata links, i think you just have to trust Emby in that regard..

[Smitty018210 159]

24 minutes ago, ginjaninja said:

Is the end goal to ensure your library metadata is right/deduped or the metadata folder on filesystem? I would be tempted to focus on getting the library metadata correct and trust that emby will sort out the filesystem (apart from wasted space, does it matter?). That wont clean duplicates on the filesystem but at least your library will be 'correct'. I dont recall the api exposes the filesystem related metadata links, i think you just have to trust Emby in that regard..

Right now I have 14,883 possible duplicate People profiles showing in my People folder. A lot of them have wrong information, duplicated information, or look like old/bad profiles that are just sitting there doing nothing.

Emby could be using a profile with outdated info, missing info, or just plain wrong info, while the correct profile/folder is sitting right next to it.

I know some of these are probably not true duplicates. Some are likely different people who happen to share the same name. I understand that part.

What I am trying to figure out is why Emby created so many of these People folders in the first place.

These are not just empty folders. They have pictures, NFO files, and profile data.

Right now I am trying to do two things:

• Remove old, bad, wrong, or duplicate profiles.
• Understand why Emby created this many People folders, so when I do not clean them up. I don't end up having the same problem happen again.

Edited 40 minutes ago by Smitty018210