Leaderboard

What It Does Aperture is a self-hosted recommendation engine that creates personalized "For You" libraries in Emby. It analyzes your watch history and uses AI to suggest movies and TV series you'll actually want to watch—complete with explanations for why each title was picked for you. Features Personalized Recommendations Per-user recommendation libraries - Each enabled user gets their own "Recommendations" library Movies and TV Series - Separate recommendation pipelines for each media type AI-generated explanations - Every recommendation includes a personalized "why you'll like this" explanation Ranked posters - Custom poster overlays showing rank and match percentage Exclude watched content - Option to only recommend unwatched titles Top Picks (Global) Popularity-based libraries - "Top 10 Movies" and "Top 10 Series" based on what's trending across all users Configurable popularity formula - Weighted combination of unique viewers, play count, and completion rate Time window - Only count watches from the last X days Auto-enabled for all users - Everyone sees the same Top Picks User Management Import users from Emby - One-click import of all Emby users Per-user controls - Enable/disable AI recommendations for movies and/or series independently Taste profiles - AI-generated synopsis of each user's viewing preferences (e.g., "You gravitate toward cerebral sci-fi and morally complex antiheroes...") Watch history sync - Delta sync support to only fetch new watches Algorithm Settings (Separate for Movies & Series) Recommendation count - How many titles to recommend per user Candidate pool size - How many candidates to consider before final selection Watch history depth - How many recent watches to use for taste profile Scoring weights: Similarity weight - How much to favor titles similar to user's taste Novelty weight - How much to favor titles different from what they've seen Rating weight - How much to factor in critic/audience ratings Popularity weight - How much to favor widely-watched titles Diversity factor - Ensure variety in genres/styles AI Configuration Embedding model - Uses OpenAI text-embedding-3-large for semantic understanding Text generation model - Selectable GPT model (gpt-4o-mini, gpt-4.1-mini, etc.) for explanations and taste synopses Cost estimator - Real-time cost estimates based on your library size and settings Library Configuration Source library selection - Choose which Emby movie/TV libraries to include STRM output path - Configure where recommendation libraries are written Library naming - Customize the names of generated libraries Job Scheduling All jobs support cron scheduling with these operations: Sync Movies/Series - Pull metadata from Emby Generate Embeddings - Create vector embeddings for new content Sync Watch History - Update user watch data (with delta sync) Generate Recommendations - Run the recommendation pipeline Sync STRM Files - Write recommendation libraries to disk Refresh Top Picks - Recalculate popularity rankings Monitoring Job run history - Track when each job ran, duration, and success/failure Real-time job progress - Live streaming logs while jobs run Database statistics - View counts of movies, series, episodes, embeddings, etc. How It Works Library Sync - Aperture syncs movie/series metadata from your Emby libraries Embedding Generation - Each title is converted into a vector embedding capturing its "essence" Watch History Sync - Tracks what each user has watched Taste Profile - Builds a vector representing each user's preferences Candidate Generation - Finds unwatched titles similar to user's taste Scoring & Selection - Ranks candidates and selects diverse top picks Explanation Generation - GPT writes personalized "why you'll like this" for each pick STRM Library Creation - Generates virtual library with STRM files, custom posters, and NFO metadata Tech Stack Backend: Node.js/TypeScript with Fastify Frontend: React with Material UI Database: PostgreSQL with pgvector for similarity search AI: OpenAI for embeddings and text generation Deployment: Docker The Result Users see personalized recommendation libraries right in their Emby home screen with suggestions that actually make sense—no more endless scrolling wondering what to watch next.

Le problème venait pas d’emby mais de mon switch ethernet. Le port que j’utilisais sur mon PCHC était défaillant. J’ai mis sur un autre port, ça fonctionne. un vrai casse-tête. En tout cas, merci pour votre aide.

All you need now is a motherboard to drop in the the 12th Gen i7 w/32GB 4800 DDR5 & Arc A380 and you can have a dedicated Emby server.

I'd really like to have 2 Factor Authentication added to the login screen. It's just this (optional) extra layer of security to help secure the server (which, especially if people use camera uploads) contains pretty private data. There are for every type of programming language quite a few libraries available, so implementation on a server shouldn't be too hard to realise

My proposal is a setting in the Emby client apps to set custom proxy headers like e.g. in Immich, Conduit or Embywatch. Currently you need to open a ton of paths like /emby/web/* or /emby/Items/* to be able to use an Emby app from the outside of LAN via Reverse Proxy with an authentication layer in front of Emby. Pangolin/Traefik Reverse Proxy is used in this example: You would set up the custom proxy header "P-Access-Token: xxx" and "P-Access-Token-Id: yyy" inside the Emby app and it can connect successfully to your Emby Server behind the Reverse Proxy with all the bells and whistles like 2FA, Passkeys etc since the Request Headers tell the Proxy: "Hey, I'm a trusted app because of your set up Headers for the resource requested but not able to use your Pangolin Login Prompt like a Browser".

Today I thought how nice it would be if I could start playing music on one device (e.g. computer #1), and then when I move around (like to computer #2, or somewhere where I only have my phone), I pause on the 1st device and then resume the now playing queue on the other device. But it appears one device has no awareness of what was playing on another device. Any chance of adding such a feature or option? I don't know how common of a use case this would be, but I've never been one to do much with music in Emby, and this kind of feature would probably make a big difference in choosing Emby for listening to my music.

Seems I had to add the discord user name. All working again now.

Hey everyone, just wanted to give a status update since it's been a month. Thank you for the feedback @Happy2Playand @Dickydodah!. It's taken a while, but after installing DrivePool, I've finally finished moving all the content to a single folder! I can't believe how much it simplified the entire process! And for anyone who thinks DrivePool is going to be very technical, let qualm concerns, it was incredibly straightforward. I think I ran into only one small issue that I figured out how to work around after a few minutes. Question. Originally, I stored the Emby install on a separate NVMe (Z:\), but for the new server, I was planning on installing it on the C drive. The actual media file and NFOs will all stay on the Z drive though. So my question for the group is, will changing the drive letter the Emby Server install is on break anything?

Yes! I can confirm that this version works with Samba shares now. Anyway, thank you for your availability and for responding so quickly to feedback. Now I can play around with these markers and home automation.

the data directory is /raid/data/module/EmbyServer/programdata @sa2000can update the doc.

But it wasn't with the F5 key, it was with the browser's refresh button, the arrow in the shape of a circle.

谢谢你，我很期待加入这个功能。M3U可以支持更多的属性标签 兼容性更多的时间变量： {utc:yyyymmddHHMMss}-{utcend:yyyymmddHHMMss} rtsp://（这是回看的IP地址）/PLTV/88888888/224/3221226103/10000100000000060000000002902003_0.smil?playseek=20260104070000-20260104182300 rtsp://（这是回看的IP地址）/PLTV/88888888/224/3221226103/10000100000000060000000002902003_0.smil?playseek={utc:yyyymmddHHMMss}-{utcend:yyyymmddHHMMss}

Can't reproduce either even with dev tools open and disable cache. Must be a browser extension that's clearing data or something.

Will check that and report back, thanks.

I am looking to see if a SAML2/OAuth plugin is something others would use. I myself would love this feature to share Emby publicly with family members and friend and worry a little less about insecure passwords that some may use. I personally am a big advocate for security as I work in IT and see how a lack of good security can cause havoc. I love Okta and what it offers. Anyone can sign up and use it for free under 1000 users if you sign up as a Developer. That is just plain awesome. Let's see who else is interested. I am excited to find out! Thanks

I would honestly pay for this...!

lol, i'll work on this one, just spent 4 hours with fiber internet issues. its off topic anyway. lol

what if you setup the same video url as a strm file?

HI, yes it makes sense. Thanks.

HI, yes it makes sense. Thanks.

Hi, thanks for following up.

And can you resume at that point?

After writing, I ran some DB maintenance tasks which seem to have brought things back to normal. The weak point with Weblate is the stats calculation (all the numbers you are seeing in the overview lists). They are cached for an hour, that's what causes the first-load-slow vs. subsequent-loads-fast effect you were seeing - and you are still seeing, but normally it's few-seconds vs. no-delay, not 30-60s on first-load. I also changed the thresholds for auto-maintenance, but not sure whether it will hit-in at the right times now. Please keep observing the behavior and let us know in case when it would become slower again. Thanks a lot for the feedback!

Bumping this as a sign of support. Just to add, even adding a new library to 5 years can be a pain in the rear. Or other config changes.

In my case now it seems to be fine, normal charging, both on the first visit and subsequent visits.

I am working on your request now to allow a admin/user to decide if they want to see the "AI WHY" I found it helpful in testing, and assume it will reduce ambiguity. In my case my kids often just watch stuff in my profile, and I was wondering how these suggestions could possibly be accurate while building the algorithm.

Aperture uses the Emby API to create libraries on the server, which are then associated with specific users. The end result is the libraries show up on any connected device that user is logged in to.

Here is a link to a package on github for the container Please read the instructions, there is a lot to understand.

It seams to work for me now, just setup a test with SMB share.

I am planning on releasing it yes! I have only been working on it for 2 days, so it is not quite ready. what I would love is some conceptual feedback and any other possibly related idea/enhancement that I could include. Be advised this is not a plugin in the traditional Emby sense. It is an external application which will run inside a Docker Container, that interfaces with Emby via the API to create libraries/collections/playlists and setup the correct library/user associations.

Figured it out. If I set as follows it works Movie Name - Part 1-thumb.jpg Movie Name - Part 2-poster.jpg Movie Name - Part 3-poster.jpg

Thanks.

That's good that you have it sorted! You can go to the Music Folders tab and rescan just the parent folder, based on your folder structure. This should pick up the changes in most instances. If this doesn't work as a first pass, then scan and/or refresh metadata on the album itself (and/or all the underlying tracks). This way you can check the changes straight away and you don't need to wait for a full library scan to complete.

Ok, I've had this problem on Synology (7.2) where my emby server is suddenly using all of my memory. I have 8Gb installed on a 1019+ system, emby server using 6Gb making system completely unresponsive. Interestingly, I have 2 other servers running - one on a debian minimal install with all the same scheduling (mine was also scanning every 12 hours). That is running at around 1.5Gb (from 32Gb installed on ryzen 7 5825U processor). This server only has 3 libraries as opposed to the 7 I have on the Synology - same libraries just use as a backup for emby server on Synology. The 3rd server I am experimenting with is a new unraid server running on a ugreen nas (12th Gen Intel® Core™ i5-1235U, 32Gb ram - I'm thinking or replacing the debian one). It has same 3 libraries as my debian instance - and that is running at 6Gb memory using all the same settings - scanning etc... All 3 servers are running the latest 4.9.1.90 emby server version with debian running from a straight docker compose file, unraid running from their app install (also docker) and Synology running from the direct download from emby themselves. I have always installed the latest stable release. This problem has only surfaced from the 4.9 install - I had no issues from the previous version and never had to restart the emby server to release the memory on Synology before. At the moment I did change the scanning settings to 1 per week on synology, twice per week on debian and left the unraid version untouched from default to see how things go though i think I might have to look at the unraid version if the memory keeps increasing. (the unraid server is only 3 weeks old at the moment which is why I'm just at the experimental stage of things) All media files are held on my Synology with mounted filesystems running on both debian and unraid. Playback from Synology has always been a bit sluggish compared to the other two systems but the Synology system is now 6 years old. If anyone is interested I can post an update as to how things are going in the coming weeks

no i was not using any tag editors, so i just installed mp3tag and yes some Album Artist were showing Beatles and not The Beatles i corrected it inside the tag editors and it is working now. Not sure if i need to rescan the librairies after editing or if it will rescan automaticalkly after a change tho. Anyway it solves the issue thanks for helping

I ended up doing that, took some trial and error. Hopefully it doesn't happen with all the disks I get. But on another note is there a way with MakeMKV to allow me make multiple MKV at the same time? It seems this tool works great with movies, but when I go to work on different tv shows, from what I have tested I can only output 1 episode at a time then change the settings to get the next. When looking up possibilities I see people asking about multiple "movies" at the same time which works completely different.

Yeah, this seems to be the root cause. I deactivated the Qsync vor H264 and will see what happens to all the other streams. By the way, it looks like that only the decoder is affected.

This is an older topic but is the first to come up in an internet search. Was having the same issue. I was hesitant to click Download because I didn't want it going straight to some device as default. However, after getting up some courage and selecting it I found out that Download will popup a normal file save window (Emby on PC). Download To is directed at downloading to registered mobile devices. As of Jan 2026 it's working properly, you simply have to use Download... not Download To.. I think the wording should be changed to Save to File and Download To instead. That wording would be clear that both options do entirely different things, in my opinion.

Thanks, I can see the issue. I'll fix it in next version. Seems one of the date values is too high/incorrect. I replaced the dateutil library as it's not a standard python library. fyi, I prepared the plugin code to make it python "no-gil" ready. I hope some day Kodi will switch to no-gil which would result in a huge performance gain (real CPU multicore support).

I had not. I will give that a try

Indeed, very cool and could be very useful for Netflix credits that last longer than 3 minutes. Also, speaking of minor or secondary things, don't forget about Samba sharing. But well done and thanks on the work already done.

Right I didn’t say it was not possible but it’s definitely not optimal for images in the web app.

Yes, the headers would need to be appended on every request. But somehow Immich was able to implement it and the app is mainly about images and videos, isn't it? These are the commits: https://github.com/immich-app/immich/commit/c2313f7a99ff0d311d616cd5c0af3d3a39896f0e https://github.com/immich-app/immich/commit/13d43e193e9fdb69428d8eece3c31d2cd6f2ca5f https://github.com/immich-app/immich/commit/fe554c3a5bb0139d874ccd34cc947c7628543e5b Regarding the web app, for this part custom headers are not needed imo since there you can login via Reverse proxy Web login form. This proposal is just for the mobile apps where using the Web Login form is not possible.

This whole clingy opening paths thing would not be necessary if Emby apps would support custom proxy headers. But this dream will never come true I guess. Feel free to give the proposal an upvote:

I told him but he hasn’t gotten back to me yet.

Major achievements? Honestly, I’m struggling to find any. From a user perspective, 2025 has mostly felt like a year of dodging critical bugs while the features nobody asked for continued to underwhelm. What I’d like to see for 2026 is for the dev team to actually enroll in a Security 101 course. It’s becoming a bit pathetic how often fundamental security best practices are ignored or treated as "low priority" while the community is left to find and document the holes ourselves. For instance, we’re still dealing with: Unauthenticated Access to Private Data: We are literally on page 6 of a thread where anyone can access library images just by knowing an ID. No token, no login, no problem. The response time on fixing basic privacy leaks is glacial. Unauthenticated access to images by itemid Permissions Bypasses: Having "Critical Security Vulnerability" in a thread title regarding Live TV permission bypasses should usually warrant an immediate emergency patch, but here we are. Critical Security Vulnerability: Live TV Permission Bypass PIN Bypasses: The fact that a simple "Home" button press can bypass security PINs is a masterclass in "Security by Wishful Thinking." Security Issue: Using home button makes PIN bypass possible And then there is the "Universal" Android app. It’s truly a "major achievement" to spend years trying to replace the perfectly functional Android TV app with a bloated, stuttering "one size fits none" mobile port. It fails miserably at providing a smooth 10-foot experience, yet the devs seem hell-bent on sunsetting the one version that actually works for TV users. If the goal for 2025 was to prove that a paid product can have slower security responses and worse UI regressions than its free competitors, then mission accomplished. Ultimately, for 2026, the devs need to start concentrating on legitimate bugs and community concerns instead of @Luketreating this like his own personal pet project. This is a paid product with a community of users, not a private sandbox where critical security and usability issues can be hand-waved away just because they don't fit a specific internal vision. Fix the foundation before adding more "achievements" to the pile.

Hi Quickmic Thanks for the new update. Have a question about the new Sync Themes option that allows you to play movie/tv theme songs. Will there be extra control added to this feature as new updates are rolled out? Reason I ask is that with TVTUNES I can tell it not to play movie theme songs and to only play tv theme songs when I enter the TV Show and view the seasons. I prefer not to move from one tv show to another and have the tv theme song play, I prefer the tv show play when I am in the tv show viewing the seasons and episodes.

Hi, just wanted to update everyone. We are in process of submitting an update to LG that should resolve this. Thanks.

+1! I’m a lifetime Emby user & I’d also be keen to run Emby on a RK3588. I can test this out once it’s available in pre-release.

This project only purpose was to show the devs that it's not forbidden that an awesome project "under the hood" like Emby can also look stunning. In visual terms default Emby is by far not in pole anymore (if it was at all at one point). CSS is cool, but who the hell uses the WebClient for watching content? It would be awesome when the CSS work from people like@fillidillwith his CSS theme would also be beneficial for all of the clients. But this will likely never happen.