TeamPiped
[Suggestion] Recommendations
Describe the feature
One of the things that made me feel Piped is incomplete is the lack of a recommendation section where you see videos from channels you don't follow but still would enjoy watching. Problem is almost always this feature is privacy invasive so i thought it maybe doesn't have to be.
How? Using Publicly Accessible Playlists could help remedy this, basically you put the videos you like watching in a playlist where the Algorithm recommends videos similar, this has the benefit of introducing you to new content but only under your own choice.
Additionally the option for Recommendations could be just turned off altogether from the settings.
Why would this be useful to add?
because finding new content on LibreTube (Frontend to Piped) can get exhausting, this will make it easier.
Concept(s)
No response
Additional context
No response
Acknowledgements
- [X] I have searched the existing issues and this is NOT a duplicate or related to another open issue.
- [X] I have written a short but informative title.
- [X] I filled out all of the requested information in this form.
I recently started to self host Piped on my local network and also found it lacking that there was no "Recommended" section on the homepage. I was using Invidious for a couple of months before moving to Piped and found this feature lacking in Invidious too. The "Trending" section in both frontends is chock full of brain rotting tripe that I'd never want to watch (I know, it's just pulled from YT so this isn't a criticism of Invidious/Piped) .
When I was still using Invidious I'd started to think about implementing a Recommended section based on the users' watch history as I noticed the recommendations on the page of a video I watched were more relevant to me. As Piped does not enable history by default, this would only be a feature that would be enabled if the user enabled history, or the recommendations could be "scraped" each time a user watches a video.
Another way to generate relevant recommendations would be to go through the latest/top N videos in a user's subscriptions and scrape the recommended sections.
Ideally, any recommendations in either the users' watch history or subscription that were repeated would rank higher in the recommendation feed according to its' frequency.
