Is anyone able to currently use the Tubifarry indexer for youtube downloads?

[ChrisMoriarty]

Currently if I add the Tubifarry indexer (without cookies.txt or generator URL) I get the following error:

2025-10-07T01:44:37.126485454Z [Info] DownloadService: Report sent to Youtube from indexer Tubifarry. Lana Del Rey - Honeymoon - 2015 [AAC 128kbps] [WEB] 
2025-10-07T01:44:40.399530586Z [Error] YouTubeAlbumRequest: Failed to process track 'Honeymoon' in album 'Honeymoon' 
2025-10-07T01:44:40.402570801Z [v2.14.5.4825] System.Exception: Failed to extract signature deciphering algorithm
2025-10-07T01:44:40.402580359Z    at YouTubeMusicAPI.Internal.Player.ExtractSigDecipherAlgorithm(String playerJs, Nullable`1 globalVariable)
2025-10-07T01:44:40.402589589Z    at YouTubeMusicAPI.Internal.Player.CreateAsync(RequestHelper requestHelper, String poToken, CancellationToken cancellationToken)
2025-10-07T01:44:40.402599411Z    at YouTubeMusicAPI.Client.YouTubeMusicClient.GetStreamingDataAsync(String id, CancellationToken cancellationToken)
2025-10-07T01:44:40.402608685Z    at Tubifarry.Download.Clients.YouTube.YouTubeAlbumRequest.ProcessAlbumAsync(CancellationToken token)

Next I went on an adventure to learn how to self host the YouTube Trusted Session Generator service. However, the repo for the Youtube Trusted Generator is deprecated, and directs you to the companion. I have it running stable, although I am unable to add the Tubifarry indexer because it gets an error:

Failed to valiate session generator service: Error connecting to the trusted session generator: Response status code does not indicate success: 404 (Not Found).

I'll be honest, I don't know much about invidious or the companion. Has anyone also had any luck lately?

[TypNull]

Sorry you're having trouble setting up the YouTube indexer. Although the session generator was deprecated quite some time ago, it continued to work, until now. I've tested it, and it appears to finally be non-functional. The companion tool operates in a completely different way, isn't compatible, and likely never will be.

If you're using the YouTube downloader and getting blocked by YouTube, you'll now need NodeJS installed and cookies set. The token will be generated internally. This setup is straightforward on Windows, macOS, or Linux, but significantly more complex when running in Docker.

What system are you currently using?

[ChrisMoriarty]

Thanks for the quick reply! I'm indeed using docker (compose) for Lidarr at the moment, specifically this image:

ghcr.io/linuxserver-labs/prarr:lidarr-plugins

Sounds like I'd need to add NodeJS to that docker image, right?

As for having cookies set, do you mean to add the cookies.txt file added to the indexer configuration within Lidarr?

Also, I'm not 100% sure I'm being blocked by youtube. Is the "Failed to extract signature deciphering algorithm" an indication of being blocked?

[TypNull]

If you're using Tubifarry version 1.8.1.0, I recommend switching to the develop branche. YouTube is in the phase of updating its API, which can cause temporary disruptions until all components in the chain are adjusted. Since Tubifarry sits at the end of that chain, it's often the last to receive fixes.

While version 1.8.1.7 may resolve the issue, it's not guaranteed. Setting cookies and tokens manually might help, but YouTube is currently overhauling its entire authentication system for all players. We're in another transition phase, so things may work, or break, unexpectedly.

All YouTube players have currently some problems, but generally still work.

[andperks6]

I also facing this issue on the develop branch, Lidarr version 3.0.0.4856 Tubifarry version 1.8.1.6. This is with setting cookies + youtube premium.

I'm happy to provide any logs to debug this or help contribute to an alternative to trusted session generator (maybe this tool inspired by it?

The biggest issue I see independent of Youtube updating its API is Tubifarry no longer has an input field for a POToken on the develop branch - which makes it impossible to input one even manually for testing.

[TypNull]

The poToken generator no longer works because YouTube changed its API. Every YouTube downloader now requires a JavaScript engine. Tubifarry needs NodeJS installed it generates the token internally. Since tokens are now track-specific and cannot be generated externally, a manual token input field is no longer necessary.
The process is actually simpler than before. You just need to install NodeJS and provide your cookies.
I plan to add a standalone nodeJS version, but if you want to contribute you can checkout the python branch. There I started a nodeJS installer that already works on Windows but not yet on Linux and Docker.

[ChrisMoriarty]

I switched to the develop branch and it didn't seem to help. Appreciate the context, sounds like I should hold off for now to preserve my sanity.

[WoolMagician]

Hello there, getting this error with cookies.txt and Node.js installed.

2025-12-14 16:32:21.0|Debug|YouTubeRequestGenerator|Generating search requests for album: 'Louis Armstrong All-Time Greats' by artist: 'Louis Armstrong'
2025-12-14 16:32:21.0|Debug|YouTubeIndexer|Downloading Feed https://music.youtube.com/youtubei/v1/search?key=
2025-12-14 16:32:21.3|Error|YouTubeDownloadRequest|Error processing download: Requiered token is null. (Parameter 'buttons[2].musicPlayButtonRenderer.playNavigationEndpoint.watchPlaylistEndpoint.playlistId')

[TypNull]

The issue with YouTube is ongoing: YouTube constantly changes things, the YouTube API package developer receives notifications when something breaks, and then I need to update the package accordingly.
I don't know the exact internals of how the YouTube API works. I've updated the package to v2.0.2, please try it again and report back.

Also, the three lines you posted aren't nearly enough to diagnose what's happening or when things go wrong. This could be my fault since the package is currently undergoing a major rewrite and I may have missed some internal changes. Or it could be YouTube's fault, they might have changed their token system dramatically, and tokens may not last long enough, requiring me to refetch them. Or the API developer might have missed something.
It's just YouTube being owned by a company with endless money. It's a fight.

[WoolMagician]

Appreciate the prompt answer, I know that smart integrations with big services can be a constant fight.
I suggest to warn the user that this feature is, at least, known to be fiddly untill proper support is implemented (if ever possible).
Keep up the good work, enjoying the plugin a lot so far :)

[TypNull]

I regularly update the warnings, but since everything worked perfectly last week, there was nothing to report. Today, however, only half of the tracks downloaded successfully, the rest returned a "Forbidden" error.

This might be due to an issue with the token allocation, but I’d need to test that theory. The Exception comes at least not from my code and the YouTube API developer (linked in the README) hasn’t reported any either. You could reach out to them to see if they can identify the source of this error and confirm whether it’s a bug or not. If yes only he can fix it. If not I could probably.