Overview

Provide a TARGET and a REFERENCE, and Matchering produces a mastered target intended to match the reference’s RMS, frequency response, peak amplitude and stereo width. That makes it a practical option for pursuing a consistent sound against a chosen benchmark; without a suitable reference, its focused approach offers less direction than a mastering workflow built around broader creative control.

Matchering 2.0 is available as a containerized web application, a Python library and a ComfyUI node, and it is integrated into the UVR5 Desktop App. Songmastr, MVSEP and Moises host it for people who want to try it without installation. The project says the algorithm uses digital signal processing rather than a neural network, and that users may use their mastered tracks wherever they want.

Key features

Reference matching is the defining capability. Matching a reference’s loudness, tonal response, peak level and stereo width can help align a target with an established sound, but the result depends on selecting a suitable reference. The algorithm is described as working well across almost all genres, especially EDM; experimental music with a very specific musical form is a less dependable fit.

Developers can use the Python library with the wider Python ecosystem or its command-line application. The library requires Python 3.8.0 or higher and a machine with 4 GB of RAM, which is a manageable baseline for many development setups but still makes this less immediate than using a hosted service. Loading MP3 files also requires installing FFmpeg separately; WAV is supported without that stated extra requirement.

The Python package is distributed under the GNU General Public License v3 (GPLv3), so its open-source availability comes with a license to consider for software integration. For users who prefer not to install Matchering, the hosted integrations provide a route to try it through Songmastr, MVSEP or Moises.

Pricing

Matchering is free. The Open-source software plan costs 0.00 USD per free and is distributed under GPLv3; no paid plans are published. This is a straightforward fit for hobbyists, musicians and developers who can use the available forms and accept the license terms, rather than buyers seeking a paid tier with a different service commitment.

Platforms

Matchering supports Linux, macOS and Windows, as well as web and self-hosted use. The containerized web application, Python library and ComfyUI node give it several routes into an existing setup, while UVR5 integration and the three hosted services offer additional ways to access it. The self-hosting option needs particular care: the maker says the web app is intended for home and in-house use, not public Internet hosting without security and scalability changes.

The maker points to Django, SQLite, Redis and the Matchering worker running in one container, non-scalable SQLite, DEBUG enabled and the absence of a production web server as reasons not to expose the application publicly as-is. That warning matters more than the convenience of a browser-based interface for anyone considering an Internet-facing deployment.

Who it's for

Matchering makes the most sense for musicians who want a reference-guided master, developers comfortable with Python or command-line integration, and users who already work with ComfyUI or UVR5. It is less suitable for experimental material whose form is unusually specific, or anyone who wants a conventional mastering environment without choosing a reference track.

Pros and cons

Pros: Free access across multiple forms, including a Python library, containerized web app and ComfyUI node, makes it adaptable to both creative and developer workflows.
Pros: Matching RMS, frequency response, peak amplitude and stereo width gives users concrete reference targets rather than an unspecified mastering result.
Cons: The Python route requires Python 3.8.0 or higher and 4 GB of RAM, and MP3 loading needs separate FFmpeg installation.
Cons: The web application is not suitable for public hosting without changes, so self-hosters must not treat its containerized form as production-ready for the open Internet.
Cons: A reference-dependent method is a narrower choice for users seeking open-ended mastering control, and the project cautions that experimental music with a very specific form may not suit it.

Alternatives

For another free route to mastering software, compare Audio Mastering Software. Readers specifically comparing AI-based offerings can browse AI Music Mastering Software, though Matchering itself uses digital signal processing rather than a neural network.

IK Multimedia ReSing is a freemium macOS and Windows option with a free tier limited to two voices, two instruments, one RVC import and no model generation; choose it for those voice and instrument tools rather than Matchering’s reference-led mastering. Tunr is another free macOS and Windows alternative; its free key is counted per machine and starts with three masters before a licence key, so it may suit someone seeking that desktop workflow.

Voxengo SPAN is a free real-time FFT spectrum analyzer plugin for macOS and Windows, a better fit when the need is spectrum analysis rather than automatic matching. StudioZIO Mastering Suite is free macOS software supporting AU, VST3, AAX and standalone use without a licence key or account, making it an option for users who want that plugin and standalone format range.

TDR Limiter 6 GE offers a personal/small business license for 60.00 EUR per once, with free updates for one user on up to five computers; consider it instead if that paid license arrangement and platform support suit your setup. DSP-Quattro is a macOS alternative with a free trial. Acustica Audio Erin Studio and FUSER are other options; FUSER costs 59.00 GBP per once and supports macOS and Windows.

Verdict

Choose Matchering if you want free, reference-based mastering and can supply a suitable benchmark, especially if Python, ComfyUI or UVR5 fits your workflow. Its strongest reason to choose it is that focused matching is available in several forms at no software cost; its strongest reason to look elsewhere is the narrow reference-led approach, compounded for self-hosters by the warning against public deployment without substantial changes.

	Matchering	ShevtoneAudio Mastering
Free plan	Yes	No
Free trial	No	No
Paid from	Free	—
Open source	No	No
Platforms	Linux, macOS, self-hosted, Web, Windows	macOS
Free plan	Yes	No
Reference track matching	Yes	—
WAV export	Yes	—
Input formats	WAV, MP3	—

Matchering vs ShevtoneAudio Mastering

Matchering

ShevtoneAudio Mastering