Welcome to the Developer Update for December 18 – the last Developer Update of 2025!!!
Today’s featured world is Cozy Christmas by Mei ʚɞ.
Announcements
New Year’s Eve in VRChat!
VRChat NYE is rapidly approaching! The forms for booths, posters, and logos are all closed – we’re now putting it all together so you have a really cool place to hang out in with your friends.
We’re pretty excited about the theme this year and can’t wait to show you it!
Big Things Coming (for VRC+)
In January, we’re going to be adding a few perks to VRC+!
First: We’ll be doubling the number of world favorites for VRC+ subscribers, taking them from 400 to 800. This has been a very, very common request for a long time!
Second: VRC+ subscribers will have a reduced name change cooldown. This is a smaller one, but it’s also something we’ve seen requested a few times – so why not throw it in?
January will also feature our largest Content Drop yet for VRC+ subscribers. So keep an eye out for that, too.
We Yeeted the Explore Section (Under Avatars)
Or rather, will be yeeting. Yote? I’m sorry. (ed: no you’re not)
We’re in the process of rolling out a change that will remove the “Explore” section from the Avatars tab. You didn’t like it. In fact, you kinda hated it there!
So, we’ve done a few things.
First, the most obvious: most users will now no longer see the Explore tab under Avatars – it’s now in the Shop. If you still do see the Explore tab in the old location, fret not: it will be fixed in an upcoming patch.
Along with this, we’ve made a few changes to the Avatars tab to make it feel more like the “old” tab. We’ve also tried to address some other feedback with the design. Hopefully it’s a much better experience now!
Creators Bug Fixes!
High-res Thumbnail Fix
We fixed a long-standing issue in the SDK where the uploads would fail if you select a very high resolution image, e.g. 4K, to be used as an avatar thumbnail when making a new avatar.
We’ve Updated the VCC!
As the SDK 3.10.x is now the recommended SDK due to recent changes to how the Avatars uploading process works, we have updated the Creator Companion to version 2.4.5, which includes SDK 3.10.x by default for any new projects.
So.. About Moderation and that “AI Ban Wave”…
We’ve seen a lot of discussion (and a fair bit of misinformation) circulating recently regarding bans, avatars, and “AI moderation.” We want to take a moment to address these concerns directly and clarify how our systems actually work.
First: There is no “AI Ban Wave”
Let’s get this out of the way: VRChat has not deployed an automated AI system that scans 3D models for NSFW geometry, textures, or features.
- No Auto-Bans: Every single ban for content is issued by a human moderator.
- We Rely on Reports: Investigations are driven by user reports. If you were moderated for content, it is because a human reviewed a report filed against you.
- We’re Handling More Reports: Some of our strongest feedback from you over the years has been that we need to action reports more effectively, completely, and quickly. We agreed, so we have spent a lot of time improving our internal tools and processes. This has resulted in more processed reports, and thus an increase in moderation actions taken. This is why you’re seeing more suspensions than usual.
Right now, we do use classification tools that utilize machine learning for 2D images (like emoji, stickers, etc) to catch some types of violations. This scanning has been highly successful, but it does not automatically suspend users.
We plan to explore more tools that utilize image, text, and other types of classifiers in 2026 and beyond. Machine learning (not generative AI!) is a critically important tool for moderation at scale. Keeping a human in the loop for moderation decisions is even more critical.
We are very aware and very sensitive to the fact that our community is (to put it lightly) suspicious of AI or AI-adjacent tools and technology. No matter what, a human always decides to swing the hammer, not a robot.
We view this type of technology as a means to alert moderators to potential issues, proactively identify “zero tolerance” offenses, and enable our moderation capabilities to scale without overburdening tens of thousands of moderators, but never to enact bans independently.
Second: Public Avatars, “Hidden” Content, and Private Avatars
There is confusion about why avatars are being banned if the NSFW parts are “hidden” or toggled off.
First, let’s approach this by avatar type: Public versus Private.
Public avatars must never contain sexually explicit material, even if that material is never exposed by the avatar Expressions, and even if that avatar is never used in public spaces. This is for a few reasons:
- We cannot ask our moderators to examine every possible combination of behaviors and states an avatar’s Expression menu might have available.
- It is possible due to shader glitches, animation issues, or other problems, that textures or mesh may be revealed even if it was never possible to turn it on.
- As an avatar uploader or creator, you cannot ensure that a public avatar is never used in public spaces.
Therefore, if a Public avatar contains geometry or textures for sexually explicit content, it is a violation, even if there is no way via the Action menu or gesture controls to reveal those textures or geometry, and/or if you warn the user not to use it in public spaces. All sexually explicit content on avatars must be removed if the avatar is public.
We know many users buy avatars rather than make them, or have others upload avatars to your account. Please ensure that you set these avatars to private if you are not fully aware of their capabilities or their content. You are responsible for what is uploaded to your account.
Finally, we know that some creators distribute their avatars by setting them to Public, but then ask their customers to keep cloning off. These avatars are not exempt from this rule. Do not distribute avatars that contain NSFW content via this “Public sharing” method. Please either remove the NSFW geometry/textures from these avatars, or do not use the “public sharing” method to distribute your avatars.
For Private avatars, it is your responsibility to not use sexually explicit features in public instances or in spaces where you cannot validate the consent of everyone present. In addition, it is your responsibility to ensure that these features never malfunction or display incorrectly to users who do not consent.
Generally, it is best practice to keep these features limited to avatars that you do not use in public spaces. That way, you can entirely avoid the issue.
Third: “Private” Instances & Malicious Reporting
We see a lot of anxiety regarding malicious reporting in private spaces. The reality is that it is necessary for Group owners, Group members, event runners, and event attendees to exercise additional caution when it comes to ensuring that the content that they employ is used in the right place, and at the right time.
In our Community Guidelines, we state an important rule: provocative behavior restrictions may not apply in Private instances, as long as everyone present consents.
It is the responsibility of the “present authority” to ensure that everyone in the instance has provided consent. The “present authority” is the instance creator, event runners, group or venue moderators, a party host, a game master, or whatever is appropriate – it’s just whoever is “in charge” of your private instance.
For example, if you ran a group or event that had this in its rules:
“If you join this group/this event, you may or may not encounter provocative activities or content, and if you join this group/event, you indicate consent to view these activities or content.”
(The phrasing and semantics are important!)
This would serve the purpose of gathering consent. In those cases, you should also run that instance or event as an 18+ Verified Group instance. As always, use your own best judgment – it is your responsibility to ensure that attendees know what they are walking into, and you should employ every tool available to let them know, and fulfill that responsibility.
We recognize that this method is brittle at best – however, it serves the purpose to help protect you from “malicious reporting,” while impressing upon you the importance of your responsibility as someone who might run a private instance or event. Thanks to various safeguards and procedures, the risk of malicious reporting is already quite low. We plan to iterate on our systems and policies over time to further reduce this risk.
Fourth: The State of Appeals
We’ve seen comments claiming that appeals are taking weeks, or that users are being banned for “no reason.”
The straight facts: our Trust & Safety team is currently clearing appeal tickets within 48 to 72 hours. We are not behind. If you are submitting an appeal, ensure you select the correct inbox. If you’ve submitted an appeal and you haven’t heard back, check your spam folder, or log into our support portal to check your ticket (and to make sure it’s in the right inbox).
Next, we often see viral posts on Twitter, Reddit, or other platforms where a user claims they were banned for “nothing” or “for no reason.” While we can’t comment on specific cases, we do investigate many of these claims internally. The reality almost always involves violations that were omitted from the public story. Don’t believe everything you read on the internet!
Finally… we acknowledge that being suspended or banned from VRChat is different. VRChat is a home for many of us, a true third place rather than simply a thing to do. Not being able to hop onto VRChat for a weekend is very different from getting a weekend ban from a first-person shooter or a MOBA. We recognize the impact that getting suspended can have, so we want to help you avoid it at all costs.
If you get suspended or banned from VRChat and believe it’s in error, you can appeal the decision at https://vrch.at/moderation. Make sure you select the appeals inbox. Write up what happened, and be truthful! A human being will look at it.
Fifth: Moving Forward
We know that the last few weeks have been confusing, and for those of you who were suspended, frustrating.
Our goal isn’t to sanitize VRChat until it has no personality. We want VRChat to remain a place where adults can be adults, creators can create freely, and communities can thrive. However, we also have to ensure that the Public face of VRChat is safe for everyone, and that our moderation scales effectively as we grow.
We hear your feedback on clarity loud and clear. We are actively working on updating our moderation messaging to tell you what type of content caused a ban, so you aren’t left guessing. We’ve also got some other changes on the way that you might learn about soon, so keep an eye out.
Thank you to everyone who wrote in, posted detailed feedback, and held us accountable. Please keep the feedback coming. We’re listening.
Conclusion
That’s it for… this year! While we don’t quite have a date locked in yet, it’s likely the Dev Update will return January 15. It could be a week later, though!
