Neural Amp Modeler (NAM) Explained: A Guitarist's Guide (2026)

What is Neural Amp Modeler (NAM)? How NAM captures work, A2 Full vs Lite, the free TONE3000 plugin, which pedals support it, and how to play NAM on your phone.

11/1/20264 min read

Neural Amp Modeler (NAM) explained, a guitarist's guide to NAM captures

If you've read about guitar modelers lately, you've seen three letters everywhere: NAM. Neural Amp Modeler has gone from a hobby project to a feature that pedal makers now advertise on the box. In this guide I explain what NAM is, how captures work, what the new A2 format changes, and the easiest ways to play through NAM today, including on your phone.

What is Neural Amp Modeler (NAM)?

Neural Amp Modeler is a free, open-source technology for capturing the sound of real guitar gear with a neural network. It was created by Steven Atkinson, PhD. Instead of an engineer hand-coding how an amp behaves, NAM learns it: you play a test signal through a real amp or pedal, record the result, and a neural network is trained to reproduce that exact sound.

The result is a small model file, usually called a capture, that you can load in a plugin, an app or a pedal. Because NAM is open source, anyone can make captures and share them, which is why there are now hundreds of thousands of them online.

Modeling vs capturing: what's the difference?

Modeling (Line 6 Helix, Boss, Valeton's built-in amps) recreates an amp's circuit in software. You get every knob, but the sound depends on how good the model is.
Capturing (NAM, IK Multimedia TONEX, Kemper profiles) takes a "snapshot" of one specific amp at one specific setting. It can sound extremely close to the real thing, but changing the knobs a lot can move you away from the original.

Many players now combine both: a capture for the core amp tone, and modeled effects around it.

How a NAM capture is made

Play a test signal through the amp or pedal (a special sweep audio file).
Record the output with a microphone on the cab, or directly from the amp with a load box.
Train the model on the recording. TONE3000 does this online and guides you step by step.
Share or load the capture in any NAM-compatible plugin, app or pedal.

A capture of the amp alone is usually paired with a cab IR (impulse response) for the speaker cabinet. Some captures include the cab and mic too ("full rig" captures).

NAM A2: Full vs Lite

In 2026 TONE3000 and Steven Atkinson introduced A2 (Architecture 2), the next generation of NAM and now the default for new captures. Its trick is that it's "slimmable": one A2 file contains two sizes.

A2-Full: maximum accuracy, for computers and devices with spare processing power. According to TONE3000 it uses about 30–40% less CPU than the older A1 format.
A2-Lite: a smaller version of the same model, designed for multi-effects pedals and other low-power hardware.

That's why you'll see pedals advertised as "A2-Full" or "A2-Lite" support. The device simply loads whichever size it can run.

Where to get NAM captures: TONE3000

TONE3000 (formerly ToneHunt) is the main library for NAM captures and IRs, with more than 700,000 tones shared by the community. In September 2026 it launched a free TONE3000 plugin (VST3, AU, AAX, CLAP, LV2 and standalone, for Mac, Windows and Linux) that lets you browse and load captures directly in your DAW without downloading files by hand.

Which pedals and modelers support NAM?

NAM support is spreading fast. Units reported to support it include:

HeadRush Prime, Core and Flex Prime (A2-Full)
Darkglass Anagram (A2-Full and Lite)
Dimehead NAM Player (A2-Full)
Hotone Ampero II Stomp and Stage (A2-Lite)
Mooer GE1000, GE300, GE100 Pro (A2-Lite)
Nux MG-30, MG-50Li, Trident (A2-Lite)
Valeton GP series, through Valeton's SnapTone conversion (see my Valeton GP-5 vs Boss GT-1 comparison)
M-VAVE MATRX, announced with native NAM A2 and TONE3000 built in (read my M-VAVE MATRX first look)

Support levels change with firmware updates, so always check the maker's current specs before buying.

How to start playing through NAM

On a computer: install the free TONE3000 plugin or the original NAM plugin, connect your guitar with an audio interface, and load a capture plus a cab IR.
On a pedal: pick a NAM-compatible unit from the list above and load captures through its editor app.
On your phone: install JamVerse, our free app that runs NAM captures, cab IRs and backing tracks on iPhone and iPad, and plug in with a small guitar interface for your phone.

Tips for great NAM tone

Match your input level. Captures react to how hard you hit them. If the tone is too clean or too fuzzy, adjust the input gain or use calibration where available.
Pair amp-only captures with a good IR. The cab is half the sound. Try a few IRs before judging the amp.
Choose captures of the setting you want. A capture of a crunch setting won't become a clean amp just by turning down a knob.
Use headphones or a flat-response speaker (FRFR), not a guitar amp's normal input, to hear the capture as intended.

NAM FAQ

Is Neural Amp Modeler free?

Yes. NAM is open source and free to use, and the TONE3000 plugin and library are free too.

What is the difference between NAM A1 and A2?

A2 is the newer architecture. It uses less CPU than A1 and comes in two sizes, Full and Lite, in a single file, so it runs on both computers and pedals.

Is NAM better than Kemper or TONEX?

They all capture real gear. NAM's advantages are that it's free, open source and has a huge community library; which sounds best is down to the individual capture and your ears.

Can I use NAM on my phone?

Yes, with a NAM-compatible app and a guitar interface for your phone.

Play NAM on your iPhone with JamVerse

You don't need a pedal or a computer to try NAM. Our app JamVerse runs NAM amp captures, cab IRs and a full pedalboard on iPhone and iPad. It also builds backing tracks from your chord progressions, with drums and bass that follow you, so you can load a capture and start jamming in minutes.

Get JamVerse free on the App Store

Keep reading

Sources: TONE3000 NAM A2 guide, TONE3000 plugin launch, Guitar.com NAM A2 support list.