Drag audio here or choose a file
v2
Max
Wyverno Lab
Log in
Sign up
Wyverno Lab
Copy the sound style
Upload two tracks: the original and the reference. The model will reproduce the reference sound style on your voice.
File to process
Clear
Drag audio here or choose a file
Style reference
Clear
Add a reference track or select a file
Upload two files to enable the button.
Copy style
Wyverno Lab
Copy the sound style
Upload two tracks: the original and the reference. The model will reproduce the reference sound style on your voice.
Add a reference track or select a file
Upload two files to enable the button.
THE FUTURE
is here
All set. Press the button to start inference.
Preparation
NoiseGate
EQ
Compressor
Done
STUDIO QUALITY
in seconds
Result
Output
0:00 / 0:08
◂))
YOUR VOICE
any style
Base Wyverno model
Wyverno SST v2 Base
Wyverno SST v2 Base is a model designed to transfer core speech and vocal processing. It analyzes EQ and dynamic processing from reference audio and reproduces its timbre and dynamic character on the target recording. The model provides a stable, predictable foundation for daily production tasks, dubbing, building vocal bases, and further creative experiments.
Wyverno SST v2 MAX
Wyverno SST v2 MAX is an advanced model that uses a comprehensive set of the most common audio effects. This architecture enables highly accurate reproduction of reference-audio processing character on your recording. The model can faithfully replicate even complex effect chains, including unconventional vocal processing and atmospheric elements, making it especially effective for high-quality dubbing, stylistic reproduction, and professional sound production.
Examples
For dubbing studios
For musicians
Wyverno / Audio Style Transfer
Sound with its own character
Wyverno is a family of models for transferring audio processing from one sound source to another. The core idea is that the model can reproduce the processing character (dynamics, timbre, spatial and stylistic traits) of a reference audio and apply it to a new signal without manual parameter tuning.
Dynamics
Timbre
Space
Mood
Signature
Origin of the idea
The concept was proposed by the Wyverno team, known for their work on Ukrainian video-game localizations and deep immersion in high-quality audio processing.
Direction
Currently one style transfer model is available.
Wyverno Speech Style Transfer
Focused on voice and allows transferring speech processing style — for example, compression character, EQ, or the overall vocal tone.
1. Listens to the reference
The model captures dynamics, timbre, and space without being tied to presets.The model captures dynamics, timbre, and space without tying itself to presets.
2. Builds a processing profile
A style is created that can be transferred between audio sources.
3. Applies the character
The new signal takes on the reference's signature while preserving its own structure.The new signal receives the reference signature while preserving its own structure.
This approach opens possibilities for automating sound design, simplifying production, and experimenting with sound without deep manual tweaking of each individual effect.
More room for experiments
Fewer manual chains and more focus on the creative sound decision.
Wyverno adapts the character without erasing the source's uniqueness.
Wyverno Lab updates
Detailed notes on releases and experiments
Read deeper update breakdowns, behind-the-scenes stories, and new Wyverno Lab features in the Hackwhiz blog.
Go to the Wyverno Lab blog