Four tube models.
One blended sound.

VMNexus brings five tube-inspired nonlinear voicings into a native music player for iPhone, iPad, and Mac. The models are trained with the Neural Amp Modeler (NAM) A2 architecture. Select up to four at once: each processes the same signal in parallel, then their outputs are averaged into one composite voice.

The complete Premium experience is free for your first two weeks.

Neural Tube Modeling with 12AX7A, 12AT7, 5654W and 5670W selected

MODEL ROUTING PARALLELOUTPUT AVERAGED

Up to fourNAM models at once

Five voicingsDistinct tube profiles

Three devicesiPhone · iPad · Mac

Two weeksFull Premium trial

Neural Tube Modeling

NAM A2 tube voicings, trained and blended in real time.

VMNexus trains its models against parameterized, circuit-inspired nonlinear tube-response targets: gain and knee behavior, asymmetry, sag and memory, brightness, DC behavior and output shaping. These are nonlinear neural models rather than sampled impulse responses. On Mac, native-rate model banks follow the source from 44.1 to 192 kHz. On iPhone and iPad, the NAM effect branch runs at a fixed 48 kHz for sustained real-time performance while the surrounding playback path remains at the source rate.

Choose from 12AX7A, 12AT7, JAN 5654W, JAN 5670 and JAN 5670W profiles. Combine up to four without pretending they are a serial amplifier chain.

VMNexus Audio on Mac with Neural Tube Modeling enabled and four tube profiles selected in the DSP panel

The complete DSP room on Mac.Multi-tube modeling, Auto Level, Tonal Balance, Smart Enhance, parametric EQ and limiter controls in a native player.

Detailed Neural Tube Modeling panel showing 12AX7A, 12AT7, 5654W and 5670W selected

Four independent A2 models. One controllable voice.Each selected NAM model receives the same signal; VMNexus sums and averages their outputs for a balanced blend.

As far as we know, VMNexus is the first consumer music player to offer real-time blending of multiple NAM-based tube models across iPhone, iPad and Mac.

NAM A2 inside a music player

Five A2 WaveNet models learn parameterized nonlinear tube-response targets and run locally in the playback DSP path — not as a separate studio plug-in and not in the cloud.

Parallel multi-tube blend

Activate up to four tube profiles together. Their independently processed outputs are averaged into a single composite tone.

Built for Apple devices

A native experience across iPhone, iPad and Apple-silicon Mac, with a shared Premium entitlement.

Your hi-res library

Play local and SMB/NAS libraries, hi-res PCM and DSD, with external DAC support and bit-perfect output when DSP is bypassed.

VMNexus Audio on Mac showing the output device picker with a FiiO K5 Pro selected at 768 kHz DSD256, Bit-Perfect mode active