Neural Amp & Cab Modeling

The single biggest disruption in guitar plugins since 2023 is Neural Amp Modeler (NAM) — an MIT-licensed framework that lets anyone capture an amp / pedal / cab to a tiny neural network model that runs in real-time inside a free plugin. The capture / model ecosystem (ToneHunt, ToneX-converted captures, GuitarML profiles) has exploded, and paid plugins like Neural DSP / Helix Native suddenly have a free competitor that sounds indistinguishable in many cases.

For DAWs to host these in see daws-overview. For other free effects see audio-effects-free. For DSP frameworks behind these see audio-dsp-frameworks.

The free / OSS neural amp ecosystem

• ★ Neural Amp Modeler (NAM) — MIT, by Steven Atkinson. Cross-platform (VST3 / AU / standalone) plugin + Python training code (neural-amp-modeler on PyPI). Capture your own amp on a flat-spec audio interface in ~1 hour, or download community models. The standard in 2024-26 for neural amp tone.
• ★ AIDA-X — GPL3, by MOD Audio. Cross-platform VST3 / LV2 / CLAP / standalone. Compatible with NAM and ProteusVST captures + its own models. Embedded-friendly (runs on Pi / Daisy / MOD Dwarf).
• ★ Proteus (GuitarML) — GPL3; plugin loads .json JSON-format LSTM captures. Smaller models, smaller download footprint.
• SmartGuitarAmp / SmartPedal (GuitarML) — GPL3; reference plugins.
• TS-M1N3 — JUCE example pedal model.

Capture / community libraries

• ★ ToneHunt — free community library for NAM .nam model files; ★ thousands of amp / pedal / cab captures contributed by users. The huggingface-of-NAM.
• Tone3000 — paid / free community; another large NAM library.
• GuitarML Tone Library — free; Proteus-format JSON captures.
• AIDA-X Cloud — model browser inside AIDA-X.
• TONE COLLECTION (NAM) — community Discord / forums.

Impulse Response (IR) / cab simulation

• ★ Cab IRs are still essential (NAM models the amp head; you pair with an IR for the cab + mic). NAM and AIDA-X both load IRs.
• NadIR (Ignite Amps) — free IR loader.
• LSP Convolver / MConvolutionEZ (Melda) — free IR loaders.
• Hesuvi / Equalizer APO IR — system-wide IR convolution.
• OwnHammer / 3SigmaAudio / RedWirez — paid commercial IR libraries.
• CabIR.eu / GodIR / Cab-Lab — IR makers + libraries.

Free amp sims (non-neural classics, still useful)

• Mercuriall Audio U530 / Spark (free editions) — free closed-source.
• Audio Assault Head Crusher / Ignite Amps plugins — many free.
• ★ ChowDSP BYOD — GPL; "Build Your Own Distortion" pedalboard plugin; modular blocks of distortion / boost / EQ; no neural net but very high quality.
• TSE BOD / X50 / R47 — free / freemium.
• Nick Crow 8505 / Wagner Sharp — legacy free.
• TPS Plate Reverb — free.

Paid commercial amp sim landscape (for comparison)

• Neural DSP Archetype / Plini / Nameless / etc. — paid (~$200/each); arguably what NAM is replacing for many home users.
• Line 6 Helix Native — paid; Helix hardware companion.
• STL Tones / Tonality / Tonehub — paid.
• IK Multimedia AmpliTube 5 — paid; broad amp library.
• Positive Grid Bias FX 2 — paid.
• Audified amp sims — paid.

Hardware that runs NAM / AIDA-X

• ★ MOD Dwarf — paid Linux audio pedal; runs AIDA-X + LV2 plugins.
• Mod Audio devices in general — Duo / Duo X / Dwarf.
• Aida DSP — Linux DSP boards.
• Headrush Prime / MX5 / Pedalboard — paid commercial multifx; firmware updates added NAM-style cloud captures (closed format).
• Quad Cortex — paid premium multifx.
• Tonex (IK Multimedia) — paid; closed neural capture format (community converters to NAM exist).
• Daisy Seed + community NAM port — see audio-dsp-frameworks.

DI workflow (the right input chain for capture)

• DI guitar / bass into a flat audio interface (Focusrite Scarlett, MOTU M2, etc. — see audio-interfaces-hardware).
• Reamp box for amp captures (Radial Reamp, Fryette LX-II).
• Load NAM + Cab IR + EQ in your DAW.
• Pair with NeuralPi community projects for Pi-based pedal builds.

Pick this if…

• Default free neural amp sim, 2026: NAM (Neural Amp Modeler).
• Free Linux / Pi / pedal-runnable: AIDA-X.
• Smaller model footprint: Proteus (GuitarML).
• Free non-neural pedalboard: ChowDSP BYOD.
• Pre-built amp library, free: ToneHunt or Tone3000 for NAM models.
• Cab IR loader, free: NadIR or MConvolutionEZ.
• Hardware pedal running OSS neural plugins: MOD Dwarf.
• Capture your own amp: flat audio interface + NAM trainer (Python) + 1 hour of patient testing.