Free Neural Network Guitar Plugins for AI Amp Modeling (2026)

November 16, 2025

The best free neural network guitar plugin in 2026 is Neural Amp Modeler (NAM). It is open source, and guitarists share free captures of real amps and pedals for it. Other strong free options are AIDA-X, Chow BYOD, Chow Centaur, GuitarML Proteus, Stompbox, ToobAmp, and guitarix.

Each plugin uses neural networks trained on real amps and pedals to reproduce their sound, including hardware imperfections. A neural model learns how a pedal or amplifier transforms the input signal, so it can capture subtle analog nonlinearities that traditional digital signal processing (DSP) models miss.

We checked every plugin in October 2026: all eight are still free to download. If you came here for a free Neural DSP alternative, start with NAM.

1. Neural Amp Modeler (NAM): The Free Neural DSP Alternative

Neural Amp Modeler is a free, open-source amp modeling plugin created by Steven Atkinson. It plays .nam model files, which are neural captures of real amps, pedals and full rigs.

You do not need to capture anything yourself to start. Guitarists share thousands of free NAM captures on TONE3000. You can also record your own amp and train a model of it.

This is why NAM is the most common answer to "free alternative to Neural DSP". Neural DSP sells polished, artist-branded plugins (with a free 14-day trial), while NAM gives you the same kind of neural amp modeling for free, with community captures instead of a fixed set of amps.

Formats: VST3 and AU for Windows 10 (64-bit) or later and macOS 10.15 or later. On Linux, use the third-party Neural Amp Modeler LV2 plugin. Stompbox, ToobAmp and guitarix below can also load NAM files.

Price: FREE and open-source (MIT license).

Download: Neural Amp Modeler users page | NAM plugin GitHub

2. AIDA-X: Universal AI Amp Modeler & Neural Effect Player

AIDA-X loads RTNeural models to emulate guitar amps, distortion pedals, and complete effects chains. The free plugin runs on Windows, macOS and Linux in CLAP, LV2, VST2, and VST3 formats, plus a standalone app.

MOD Audio and AIDA DSP developed AIDA-X together. The project started on Arduino-based audio shields and grew into a flexible model player for the MOD Audio ecosystem.

AIDA-X is built on RTNeural, with FFTConvolver for cabinet impulse responses. Its combined models are trained at multiple gain settings, so you can adjust the gain of the neural model in real time.

Price: Completely FREE and open-source.

Download: AIDA-X GitHub | MOD Audio Page | AIDA DSP Plugin Bundle Github

3. Chow BYOD (Build-Your-Own-Distortion): Custom AI Pedal Creator

This free guitar VST lets you build custom distortion effects by chaining 43 modules, including drives, tone shapers, and utilities. It works like a virtual pedalboard for designing your own overdrive.

It started in 2019 with Jatin Chowdhury's class project for Music 420, a tape emulator plugin. His work has since grown into ChowDSP, which develops audio plugins such as BYOD.

BYOD combines RTNeural-based neural components with traditional DSP circuit modeling, with low aliasing and accurate analog behavior.

Price: FREE and open-source (GPL license).

Download: ChowDSP Website | GitHub Repository

4. Chow Centaur: Neural Network Klon Centaur Emulation

Chow Centaur is a digital recreation of the Klon Centaur overdrive pedal, which sells for $3,000+ on the used market. MIT-trained engineer Bill Finnegan hand-built the pedal starting in 1994, and it has become the "holy grail" of overdrive.

The plugin gives you the Klon's transparent overdrive: it keeps the character of your guitar and adds warmth and harmonics.

It has two modes: Traditional (circuit modeling with nodal analysis) and Neural (a real-time recurrent neural network running on RTNeural). The neural model captures the Klon's famous circuit that doubles the 9V supply to 18V.

Price: FREE VST3, AU, and LV2 plugin for desktop and iOS.

Download: ChowDSP Guitar Pedal VST | GitHub

5. GuitarML Proteus: Capture Your Own Guitar Amp & Pedals

Proteus lets you create neural models of your own guitar amplifiers and effects pedals. The free plugin supports "knob captures" (a knob's full range) and "snapshot captures" (a single setting).

Engineer Keith Bloemer launched GuitarML in 2020 to make amp modeling accessible to everyone. After three years of work on network architectures, Proteus became a free alternative to Kemper and Neural DSP for capturing your own gear.

Proteus runs LSTM networks on the RTNeural engine. A 40-unit LSTM model uses about 2% CPU, on par with commercial plugins.

Price: FREE with optional Patreon support.

Download: GuitarML.com Free VST Plugins | Proteus GitHub

6. Stompbox: Complete AI Guitar Pedalboard

Stompbox simulates a full guitar amp and effects pedalboard with Neural Amp Modeler support, cabinet IRs, time-based effects, and MIDI control. It runs headless on a Raspberry Pi for DIY hardware builds.

Mike Oliphant built Stompbox to run both as Raspberry Pi hardware and as a DAW plugin. It includes a tuner, a looper, and a full effects chain.

Stompbox combines Neural Amp Modeler and RTNeural with algorithmic DSP effects, and offers remote UI and VST3 modes.

Price: FREE and open-source.

Download: Stompbox GitHub | VST3 UI

7. ToobAmp: Raspberry Pi Neural Guitar Effects

ToobAmp is a set of 22 LV2 plugins optimized for Raspberry Pi, including TooB Neural Amp Modeler, ML amp models, cabinet IRs, and classic guitar effects. It suits DIY guitar pedal hardware projects.

Robin Davies created ToobAmp for the PiPedal project. It shows that neural amp models can run efficiently on a $35 computer, which puts AI guitar effects within reach of hobbyists. If you build on an NVIDIA Jetson Nano instead, note that it has no analog audio output: set up a Raspberry Pi I2S sound card on the Jetson Nano first.

Its RTNeural and NAM models are optimized for ARM processors, so they deliver professional guitar tone on low-power hardware.

Price: FREE and open-source.

Download: ToobAmp GitHub | PiPedal Project

8. guitarix: Linux Amp Simulator With Neural Models

guitarix has been a staple of Linux audio since the 2000s. Version 0.45.0 added machine learning amp modeling, and it now loads both NAM (.nam files) and RTNeural (.json/.aidax) models. It also offers modular preamp and amp models, effects, and cabinet emulation.

Because RTNeural sits alongside two decades of DSP development, you can mix neural amp models with traditional circuit models in the same rig.

Price: FREE and open-source (GPL).

Download: guitarix GitHub | LinuxMusicians Forum

How to Make Your Own AI Guitar Effects

The easiest way to make your own neural amp model is to capture real gear with NAM or Proteus: you play a test signal through your amp or pedal, record the result, and train a model on the pair.

To build your own plugin, most tools in this list use RTNeural, an open-source inference library for real-time audio. Jatin Chowdhury created it to run neural networks with low latency and low CPU usage, which makes neural guitar effects practical in the studio and on stage. NAM uses its own open-source core library instead.

RTNeural supports several backends (Eigen, xsimd, and plain C++ STL) and layer types including Dense, GRU, LSTM, and Conv1D. These are well suited to modeling the nonlinear behavior of tubes, transistors, and analog circuits.

For neural models that generate or transform whole sounds rather than model an amp, see our guides to RAVE by IRCAM, DDSP by Google Magenta and AFTER, IRCAM's MIDI-controllable diffusion model. You can train a RAVE, DDSP or AFTER model on your own recordings without a local GPU; the FAQ explains how much audio each model needs.

Larger neural models, such as audio restoration and stem separation, are too heavy for a real-time guitar plugin. The Neural Analog VST runs them in the cloud, so you need no GPU.

If you also want free instrument plugins, sforzando plays free SFZ sample libraries such as pianos, drums and strings.

Free Neural Guitar Plugins FAQ

What is the best free alternative to Neural DSP?

Neural Amp Modeler (NAM) is the closest free alternative. It uses the same kind of neural network amp modeling, it is open source, and the community shares free captures of real amps and pedals on TONE3000. Neural DSP plugins are paid, with a free 14-day trial.

Which free neural guitar plugins work on Linux?

guitarix, ToobAmp, AIDA-X, Chow BYOD, Chow Centaur and GuitarML Proteus all have Linux builds. For NAM captures on Linux, use the third-party Neural Amp Modeler LV2 plugin.

Can I capture my own amp for free?

Yes. Neural Amp Modeler and GuitarML Proteus both let you record your own amp or pedal and train a neural model of it, then play it back in the free plugin.