A Note to Our Community: AI Studio & the Future of Reallusion 3D

iClone Official
1

Hi everyone,

Storytelling is at the heart of everything we create. While the medium evolves, the power of storytelling remains timeless. As we launch AI Studio, we have seen incredible support from many creators, but we also recognize the valid concerns expressed by our community, and we want to address them openly.

We see the profound impact of AI, but we firmly believe in the irreplaceable value of the traditional 3D craft. That is why Reallusion AI Studio was born: to bridge traditional CG workflows with generative AI, giving artists the best of both worlds.

1. AI as an Assistant, Not a Creator

Traditional 3D is the foundation of great work. AI Studio is built to expand your toolkit, keeping you in the driver’s seat to decide exactly how AI fits your creative process.

Through iClone and Character Creator, you can design characters, direct performances, and build structured 3D scenes, using AI to rapidly visualize ideas while maintaining total control over motion, cameras, lighting, and composition. Whether you build everything by hand or accelerate your workflow with AI, the choice is entirely yours. Characters gain continuity, scenes gain cinematic direction, and the technology assists your workflow without replacing your artistry.

2. Full Commitment to Traditional 3D & RTX Render

Our dedication to core 3D tools remains unchanged and is our top priority. Everything announced in our “What to Expect in 2026” preview is absolutely still coming. This includes the highly anticipated RTX Render free update developed with NVIDIA, bringing path-traced realism directly into iClone and Character Creator. We are continuing to heavily invest in the core tools you rely on every day.

3. Why Cloud Processing is a Must

To be completely straightforward: the highest-end, most powerful generative models, such as Seedance 2.0, Nano Banana, Veo3 and Kling3 are built specifically to run on cloud. Shifting to a cloud infrastructure is a must if we want to give all of our users equal access to these cutting-edge tools, regardless of local machine specs.

This approach completely removes the hardware bottleneck and eliminates the need for expensive upfront computer upgrades. Furthermore, it ensures we can continuously update and integrate the newest, most advanced AI models as the technology evolves.

4. Transparency on Cloud Pricing

The decision to leverage cloud infrastructure directly influences our pricing structure. Running, hosting, and maintaining these massive, state-of-the-art models requires substantial, continuous server infrastructure.

To keep these high-performance servers operating at the highest level and cover live processing costs, we do need to charge for generations. This allows us to offer a flexible, per-generation usage model instead of demanding hardware upgrades, while ensuring you always have access to the latest, most powerful AI technology.

The Future of Storytelling

Our ultimate vision is to champion the powerful combination of traditional 3D and generative AI. By blending the precision of 3D production with the speed of AI generation, we are delivering a hybrid productivity workflow that provides the best of both worlds. This gives you a clear path to make the most of both technologies on your own terms, keeping your unique artistic voice at the center of the creative process. Thank you for sharing your thoughts and helping us shape the future of our tools.

5 Likes
3

The integration of local usage for models like LTX or WAN should be possible, as before with the AI render beta.

Cloud usage will become either much more expensive in the future or usage limits may be introduced or models might get shut down, as was Sora. It’s possible to do a lot more than somebody might think, even with an 16 GB VRAM GPU and 32 GB RAM.

And don’t forget, that Nvidia plans to introduce real-time, GPU accelerated AI filters with DLSS 5. This could replace a lot of the current AI Studio functionality.

I don’t think that gigantic Cloud AI models will be the future. It’s too expensive in the long run. Specialized AI models, which only have limited hardware requirements and can be run locally, will become good enough to replace cloud models in the long run.

2 Likes
4

Yup local models like Infinite Talk etc should be available locally like Ai render beta.

5

The Future of iClone – Thoughts from an Indie Filmmaker

After spending several days reviewing and thinking about the new AI Studio update, I wanted to share some thoughts on iClone’s current direction and where I personally hope Reallusion focuses its future development.

First, I want to say that I understand why many users are excited about the recent AI developments. There is a lot of impressive technology being introduced, and I genuinely appreciate Reallusion continuing to push innovation forward.

However, speaking as someone who has used iClone and CC for several years as part of a serious indie CGI film production pipeline, I find myself increasingly concerned that Reallusion may be drifting away from what originally made iClone such a valuable tool.

What sold me on iClone was never the idea of a fully automated end-to-end filmmaking solution. What sold me was this:

  • Fast and flexible character animation
  • Tight integration with Character Creator
  • Rapid iteration for cinematic storytelling
  • The ability to integrate with other specialized tools such as Unreal Engine, Resolve, Blender, and external asset pipelines

I want to maintain artistic control over my productions:

  • I want to build and customize my own sets
  • I want full control over scene lighting
  • I want to choose the rendering pipeline that best fits my project
  • I want the freedom to combine tools instead of being locked into a single ecosystem

Because of that, I personally have little interest in AI systems that attempt to automate the creative side of filmmaking.

What I desperately need instead is AI assistance with the technical grind of animation production.

That is where I believe Reallusion has an enormous opportunity.

Where AI Could Truly Transform iClone

1. Animation Cleanup and Refinement

This is, in my opinion, the single most important area for AI integration.

Examples:

  • Fixing foot sliding automatically
  • Smoothing root motion transitions between clips
  • Correcting stair movement and uneven terrain interaction
  • Improving hand positioning and gesture cleanup
  • Stabilizing motion after baking layered animations
  • Automatically refining body weight and balance

These are the kinds of repetitive, time-consuming tasks that consume enormous amounts of production time.

One feature that would be incredibly powerful:

  • defining two keyframes and allowing an AI assistant to intelligently animate the motion between them based on a text description or emotional intent.

2. Animation Blending and Interpolation

Animation blending in iClone can still become frustratingly manual, especially when mixing clips generated through different workflows.

AI-assisted blending could help:

  • merge multiple performances naturally
  • preserve body weight and momentum
  • create style variations (limp, shuffle, exhaustion, hesitation, etc.)
  • add subtle behavioral quirks to characters automatically

For example:

  • habitual eye twitches
  • facial grimaces
  • nervous gestures
  • idle behaviors

These are the small details that make characters feel alive.

3. Lip Sync Improvements

This is another area where AI could dramatically reduce workload.

AccuLips was a great addition, but compared to recent advances in AI-driven facial animation, the system now feels overdue for another major evolution.

Tasks that currently take hours could potentially become much faster:

  • phoneme refinement
  • emotional facial timing
  • subtle mouth corrections
  • more natural transitions

And there is also a very practical workflow issue:

AccuLips still does not support WAV audio directly.

Because of this, I currently must:

  1. convert dialogue to MP3 for iClone,
  2. animate using the MP3,
  3. export the resulting audio,
  4. then manually replace or splice it with the original WAV files later for final production quality.

That process becomes extremely tedious on large productions.

As such, I am very excited to see what is coming with the next release of AccuLips!

4. Motion Director and Crowd Systems

Motion Director and Crowd Simulation are promising systems, but they still require a large amount of manual work.

AI assistance here could help with:

  • path planning
  • collision avoidance
  • natural idle behaviors
  • reactive movement
  • group coordination
  • automated environmental interaction

5. Unreal Engine Pipeline Stability

This is a major one for filmmakers using Unreal Engine.

Preparing an iClone scene for transfer to Unreal can become extremely fragile and time-consuming:

  • baking constraints
  • managing sequencers
  • verifying transfer settings
  • handling repeated transfers
  • troubleshooting Live Link issues
  • dealing with transfer errors such as “Wait for the previous operation to complete”

What I would love to see is:

  • dramatically improved transfer reliability,
  • smarter transfer management,
  • automated troubleshooting,
  • and ultimately a much more seamless bridge between iClone and Unreal.

Personally, I believe Reallusion’s greatest strength is not competing with Unreal or Blender as a renderer.

Its greatest strength is enabling rapid character animation and cinematic iteration.

That is why I use it.

Final Thoughts

I’m not opposed to AI integration in iClone.

Quite the opposite. I simply believe the most valuable role for AI is not replacing the filmmaker — but assisting the filmmaker. I don’t need an AI movie generator.

What I need is an intelligent animation collaborator:

  • something that helps solve technical friction,
  • accelerates iteration,
  • improves motion quality,
  • and frees me to focus on storytelling, cinematography, pacing, atmosphere, and performance.

That, in my opinion, is where iClone could become truly extraordinary.

8 Likes
6

As I’ve stated in previous posts, Reallusion has to appeal to future users of its software ecosystem.
as well as current users who have already made a huge substantial investment in NVIDIA GPU hardware.

I have yet to see any AI images or Animations ,(run on someone’s local PC Using comfy UI etc) that comes close to what we see people doing with seedance2 ,kling or even grok
( if you have some examples that prove otherwise please share).

Now understand this is not to say that there is no future for people wanting to run AI Systems on their local PC Hardware accept that I really do not believe that this will be done on the current, nor any future , generation of RTX GPU’s
I think the future of local AI computing Has pretty much been decided by NVIDIA when they introduce this new piece of purpose-built AI hardware.

7

That’s just an ARM processor with CUDA cores directly integrated, so you no longer need a dedicated Nvidia GPU. That’s very useful, because you can use your full system RAM for loading large AI models. Currently, spilling from VRAM to system RAM will slow down interference by an order of magnitude on x86 processors. With this this chip, the slowdown will not happen, although it is, of course, much slower than a dedicated GPU.

Alternatively you can already do AI interference on Snapdragon or Apple M processors, they just don’t have CUDA acceleration but use libraries like Metal. Nvidia’s chip is just a third alternative and CUDA is often a requirement for AI acceleration.

8

This here is pretty impressive:
https://www.reddit.com/r/StableDiffusion/s/siTlSGJYxZ
https://www.reddit.com/r/StableDiffusion/s/MUy8Y5EFer
https://www.reddit.com/r/StableDiffusion/s/f47pbod8Uj

14

If Reallusion is providing AI render both at local and cloud from begining then they will cater to all use cases.

Believe me local generation is as much important as cloud.
AI Studio Local Mode and RTX rendering will make core RL product a market standard.

15

I agree with most everything you mention, particularly that the most useful things AI could do would be the repetitive and motion-correction tasks that take time away from the creative aspects of animation and storytelling. I’d like AI to be an assistant, not a partner.

3 Likes
16

All that your saying is true, but through commercialize of this product, no matter what is said, it won’t be enough to stop this momentum toward (AI). The choice they’re making, makes sure our voice will be heard less. Over the bottom line of two controlling the narratives that are at play personal information, and data collection & money. How much is a content creator willling to spend? And with the provision of youtube algorithm and parameters how will this help and or address our issues?
Monies spent by wayward artist, time, and effort on products retailing unchecked value by analgorithm calling work (AI slop). So yes it’s personal for all of us to ask the question they didn’t:
Who does it help?
What’s the reasoning behind the user not having a clear choice payment of (AI)?
Where does it fit into daily work without needing to be online?
When do start understanding core of reallusion is fading?
Why will I need a cloud server account?
How can it benefit us?
On what used to be free and fun to generate characters from scratch
that we’ve been enjoying, is now but one update away from existence.

2 Likes
18

question as an outsider using iclone 7 - ( may also help RL refine their marketing pitch )

what advantage do i have using iclone 8 and AI studio over simply rendering a base image in iclone 7 and buying 660 credits for $10 at kling to animate it ?

1 Like
20

It can save time and make the workflow simplified if you have a built-in integration of these tools in your animation software. This AI Studio was made with this goal in mind, a general online platform for AI video generation wasn’t.

Also things like making character sheets for consistency are also simplified if they are all built into a single software instead of having to spread your workflow to many different platforms.

I think management is a big advantage, even ignoring all the other benefits.