Metadata for audiobooks are not fecthed anymore (used to work)

[mrjay42 5]

Hi all,

 

My problem

My Audio Books' metadata are not fetched from the internet/detected/retrieved from the metadata "service".

So lately, ALL my newly added audio books look like this:

(You can probably notice that the audio books are in French, maybe the problem is related to this? My server's display language is English, some of the content available on it is in French. That doesn't cause any problem for the other types of content: films, TV series, etc.)

 

Older ones ("older" as in added on my Emby before a certain date (undefined right now, is it before a particular update of Emby? a plugin?) look like that:

You can see that most of them are/were correctly detected.

 

 

What I have tried

• Restarting Emby
• "Forced" refresh metadata
• "Forced" scan library files"
• Fidgeting in the options -> didn't find anything significant, that's actually the first reason I'm posting here
• Tried to identify manually, and this the most interesting -> when I type the correct title of an audio book -> emby immediately displays "No items matching your search criteria were found." -> it doesn't really search, it feels like there's no http request to any API or anything happening. (see screenshot below) ----> this is actually the second main reason for me posting this here.
• Pulling Emby's image (I'm using docker compose, with Portainer)

 

 

Software Setup

• I self host my Emby server on a PC in my home, but I access it through the internet via reverse proxy (see the docker compose below if you're interested)
• OS: Linux Debian Stable - Debian GNU/Linux 13 (trixie)
• Emby: Version 4.9.3.0

Docker Compose stack

version: "2.3" 

services:
  emby:
    image: emby/embyserver
    container_name: embyserver
    #runtime: cpu # Expose NVIDIA GPUs
    networks:
      - private_secured_network
    #network_mode: bridge # Enable DLNA and Wake-on-Lan
    environment:
      - UID=1000 # The UID to run emby as (default: 2)
      - GID=1000 # The GID to run emby as (default 2)
      - GIDLIST=1000 # A comma-separated list of additional GIDs to run emby as (default: 2)
      - NVIDIA_VISIBLE_DEVICES=all
      - NVIDIA_DRIVER_CAPABILITIES=compute,utility,video
    volumes:
      - /home/bla/Desktop/apps/emby:/config # Configuration directory
      - /home/bla/Desktop/data/Downloads/transmission/completed:/mnt/share1
    ports:
      - 8096:8096 # HTTP port
    #  - 8920:8920 # HTTPS port
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
    restart: always
    labels:
      - "traefik.enable=true"
      - "traefik.http.routers.emby.rule=Host(`jayflix.bla.stream`)"
      - "traefik.http.routers.emby.entrypoints=websecure"
      - "traefik.http.routers.emby.tls.certresolver=myresolver"
      - "traefik.http.services.emby.loadbalancer.server.port=8096"

networks:
  private_secured_network:
    external: true

But that shouldn't matter for the problem at hand.

Plugins

I don't know if this is of any importance, but this is a screenshot of the installed plugins

Logs

I have attached the latest logs from my Emby server

embyserver.txt

[Luke 42077]

Hi there, what metadata fetchers are enabled on the library?

[mrjay42 5]

On 20/01/2026 at 03:57, Luke said:

Hi there, what metadata fetchers are enabled on the library?

Ok that seemed to be the issue!

I Did this ->
Connect to Emby (as admin)
-> Gear icon top right corner
-> Library
-> click on the three dots for my "Audiobooks" library
-> click on edit
-> on the panel that opened, I could see that ALL metadata fetchers were...deselected for some reasons. So I enabled them all. Right now I'm refreshing the metadata, this is gonna take a while I guess. I'll come back here to say if it worked and pass this topic as "solved"

Thank you @Lukefor your help

[mrjay42 5]

Ok so it's kinda working now, but far from being perfect.

What do I mean:

Some of the audiobooks are being identified correctly.
But weirdly, some EXTREMELY well known audiobooks are not identified. For instance, "La servante Ecarlate" (The Handmaid's Tale) is NOT identified, and even manually, if I type "servante" in the search tool, I get less well-known audiobooks, but not the "La servante ecarlate".

 

So, now I want to check with you whether my Audiobooks metadata fetchers are 'correct' and/or they could be changed/improved.

This is what I have:

 

 

 

 

If you see anything wrong here, please tell me.

Also if there are other "metadata fetchers" that I should install ->  do not hesitate to tell me, I'll install them

Thanks again for your help  @Luke

[Luke 42077]

On 1/21/2026 at 8:50 AM, mrjay42 said:

Ok so it's kinda working now, but far from being perfect.

What do I mean:

Some of the audiobooks are being identified correctly.
But weirdly, some EXTREMELY well known audiobooks are not identified. For instance, "La servante Ecarlate" (The Handmaid's Tale) is NOT identified, and even manually, if I type "servante" in the search tool, I get less well-known audiobooks, but not the "La servante ecarlate".

 

So, now I want to check with you whether my Audiobooks metadata fetchers are 'correct' and/or they could be changed/improved.

This is what I have:

 

 

 

 

If you see anything wrong here, please tell me.

Also if there are other "metadata fetchers" that I should install ->  do not hesitate to tell me, I'll install them

Thanks again for your help  @Luke

HI, you could also try OpenLibrary from the plugin catalog. It sounds like you might want to disable musicbrainz and audiodb.

[mrjay42 5]

7 hours ago, Luke said:

HI, you could also try OpenLibrary from the plugin catalog. It sounds like you might want to disable musicbrainz and audiodb.

Ok I'm gonna try this!

I'll update this post when I'll have some new results

[brothom 177]

1 hour ago, mrjay42 said:

Ok I'm gonna try this!

I'll update this post when I'll have some new results

Recently I got some of my audiobooks digitized and found that this setup works fine for me: 

It's important to have the audio metadata itself also written correctly, especially Artist, Album and Track (for the correct order).

I used mp3tag to apply the correct metadata to the audiobook chapter I'd digitized myself as shown below:

 

[Luke 42077]

I think it depends on your content. If none of your audio books are on musicbrainz or audiodb, then you should disable those to avoid possible false matches.

[mrjay42 5]

On 23/01/2026 at 15:29, brothom said:

Recently I got some of my audiobooks digitized and found that this setup works fine for me: 

It's important to have the audio metadata itself also written correctly, especially Artist, Album and Track (for the correct order).

I used mp3tag to apply the correct metadata to the audiobook chapter I'd digitized myself as shown below:

 

I'm gonna try this

Because for now the results are 'meh'

[Luke 42077]

Please let us know how things go. Thanks.

[mrjay42 5]

Ok so here's what I did:

 

• As suggested by @brothomI have installed/enabled the following plugins "TheAudioDB" and "OpenLibrary"
• I selected them as the metadata manager for my library called "Audiobooks"
• And I went to my favorite website and downloaded the most well-known and most available audiobook, which is "Dune - Frank Hebert", in French. Because my "audiobooks" library is ONLY in French
• And all the metadata were retrieved perfectly and immediately for THAT BOOK

BUT
I have so many other books that are NOT identified!! Even when I try to do it manually:

 

For instance, you can see in the bottom row of the above screenshot, books by "Pierre Lemaitre" who is, without any doubt, a famous author (https://en.wikipedia.org/wiki/Pierre_Lemaitre ) in wikipedia page is literally translated in THIRTY languages. This is not an old obscure random unknown author.


And when I type HIS NAME, in the search bar, to identify manually one of his books, I get this:

Literally nothing.

And when I search for the book title "sacrifices", I obviously get a long list but NO book by Pierre Lemaitre.

 

So, the questions remain:

• Did I do anything wrong?
• Is there some kind of metadata manager/fetcher that is adapted to French authors/books/audiobooks?
• Did anything change recently in Emby that could cause this problem to emerge?

Thanks again to all of you  for your help
Thanks @Luke@brothom

 

[brothom 177]

@mrjay42you're getting there!

For Emby to correctly identify a book it also needs the metadata, see my previous explanation: 

On 23/01/2026 at 15:29, brothom said:

It's important to have the audio metadata itself also written correctly, especially Artist, Album and Track (for the correct order).

I used mp3tag to apply the correct metadata to the audiobook chapter I'd digitized myself as shown below:

 

In the previous example I had to correct the author J.K. Rowling, as something had put "I.K. Rowling" in and the book name (it didn't include Harry Potter and).

The "Album" and "Artist" metadata are most important here.
Using these two fields separately, Emby will be able to figure out which book you're actually referring to.
Combining both fields, it allows Emby to group the underlying media files in a sort of playlist style.

 

20 minutes ago, mrjay42 said:

I have so many other books that are NOT identified!! Even when I try to do it manually:

If you want to automate the process, or have an application try to identify your audiobooks for you there are two popular options:
https://www.audiobookshelf.org/ and https://calibre-ebook.com/. One is a separate server and (I think) the other is just an application.

I also had a lot of issues identifying books, especially books that are slightly less popular (I'm looking at you Goosebumps!).

---

 

Regarding Pierre Lemaitre's Sacrifices, it does exist on openlibrary.org: https://openlibrary.org/works/OL23231222W/Sacrifices?edition=key%3A/books/OL32370833M.

What you could do, is if manually identifying (via search) doesn't give you the right result, is enter the ID manually in Emby via the Edit function. That should basically always strictly identify the media:

 

Edited January 27 by brothom

[mrjay42 5]

@brothomthanks again.

So the manual identification worked!
Thanks.

It's not automatic, but at least fully manual identification via Open Library Work ID works