Version 9 -- Added "Reinforced Learning"
(TRAINING VIDEOS BELOW)
The first ever VaM script to be 90% edited by AI (ChatGPT). Originally started by a VaM’er who built the basic framework but never finished it, the script was then completely re-written by AI—making mutation creation over 100x faster and more efficient.Working alongside AI, I added the new functionality I always wanted—and hopefully it’s what you’ve been waiting for too.
Turn new character creation into a fast, visual workflow from all your existing morphs!
NovaVR Morph Mutations lets you generate random morph sets, bias them toward your favorite looks, and capture clickable thumbnails — all from an in-scene canvas so you never leave play mode. Built for creators who want speed, control, and zero clutter.
- In-scene workflow: Tweak, test, and capture without opening extra panels.
- Fast iteration: Bias toward keepers, bookmark them as thumbnails, and revert if needed.
- Clarity: Every control is labeled and reachable on the canvas, so you stay in flow.
Key features
- One-click Randomize: Generate a fresh mutation instantly with Generate Random Mutations on the scene canvas.
- Live Morph Filter: Type a substring to include/exclude morphs on the fly (title + input live on the canvas). Add a negative sign in front of the text for another way to filter.
- Clickable Catalog Grid: Auto-capture thumbnails of each look; click any tile to re-apply that exact morph set.
- Reset Tools on Canvas:
- Reset Catalog — clears the thumbnail grid and starts fresh.
- Reset Bias — wipes all bias weights back to neutral (x1.00).
- Bias System (Fine-tuning your looks):
- Add Bias (+100%) to nudge RNG toward what you like. Making that mutation and its morphs 2x more likely to appear in the next mutation. You can train the RNG to what you like best!
- Show Bias Summary to see top weighted morphs.
- Three buttons appear allowing you to apply the top morphs to your model based on bias.
- Reinforced Learning
- Brand new in Version 8. Only one slider added and the software does the rest!
- Reinforcement Learning lets a system learn by trial and feedback from the user
- Basically it uses your bias selections repeats a portion of that bias for all new mutations! Cool!
- See the help document in the .VAR for more explanations and use
- Lightweight, offline, VaM-native: Uses standard VaM UI prefabs — drops in cleanly and plays nice with existing scenes.
It's free. My contribution to all the great modelers here who have posted free content. Thank you!
