A powerful macOS companion for Neural Amp Modeler (NAM). Streamline your workflow by building complex processing chains, batch reamping audio, and training neural models—all in one place.
This project is built upon the incredible work of Steven Atkinson (@sdatkinson), the creator of Neural Amp Modeler (NAM).
NAM Reamp Lab uses the open-source neural-amp-modeler Python package for its training backend. We are deeply grateful to Steve and the NAM community for making high-quality neural modeling accessible to everyone.
NAM Reamp Lab is designed for tone creators who want to capture their perfect "digital" signal chains into portable NAM models. Whether you're stacking a boutique overdrive into a legendary amp sim and an IR-based cab, NAM Reamp Lab automates the tedious parts of the process.
- Build - Stack any AU plugins (pedals, amps, cabs) in custom chains.
- Toggle - Select which chains to include in your batch process with a single click.
- Process - Reamp your DI signal through all selected chains automatically.
- Train - Launch batch training jobs using the high-performance WaveNet architecture.
- Universal AU Support: Use any 3rd-party Audio Unit plugins (amp sims, delays, compressors).
- Interactive Toggles: Mark specific chains for batch processing directly in the sidebar.
- Deep Preset Management: Plugin states are captured automatically during reamping.
- Real-time Monitoring: High-performance audio engine with low latency.
- One-Click Batching: "Process & Train" takes you from raw audio to a training queue instantly.
- GPU Accelerated: Native support for Apple Silicon (M1/M2/M3) via Metal (MPS).
- Quality Analysis: Real-time ESR tracking with color-coded quality bands.
- Automated Export: Models are named and saved to your library automatically.
- Multi-Channel Support: Selective input channel routing for professional interfaces.
- Visual Feedback: Precise RMS metering with vDSP acceleration.
- Stable Switching: Atomic device configuration to prevent CoreAudio crashes.
- macOS 15.0+
- Python 3.10 (specifically 3.10 for stable MPS support)
- NAM Audio Unit Plugin (highly recommended)
-
Clone the Repo
git clone https://github.com/profmitchell/NAM-Reamp-Lab.git cd NAM-Reamp-Lab -
Set up Python Environment We recommend using a dedicated virtual environment:
python3.10 -m venv .venv source .venv/bin/activate pip install torch torchvision torchaudio pip install neural-amp-modeler -
Open in Xcode
open "NAM Reamp Lab.xcodeproj"
- AudioEngine: A modular wrapper around
AVAudioEngine, handling real-time routing and offline rendering. - NAMTrainer: Manages detached Python subprocesses for training, ensuring the UI remains responsive.
- ChainManager: Oversees persistent storage and deep-copying of plugin states.
The engine is divided into specialized extensions:
AudioEngine+Plugins: Complex AU/NAM state management.AudioEngine+Metering: High-performance vDSP-based RMS calculation.AudioEngine+Devices: Robust hardware discovery and mapping.
Distributed under the MIT License. See LICENSE for more information.
We welcome contributions! Please feel free to open issues or submit pull requests to help improve the NAM ecosystem.
Made by Mitchell Cohen in collaboration with the NAM Community.