Neiro

Overview

Neiro keeps audio processing on the user's machine, with a Tauri desktop app and a browser interface launched using neiro ui. The core digital signal processing tools work without downloaded models, so users can start with that foundation before deciding whether optional neural backends are worth adding.

This is more than a restoration utility: source separation, music transcription, editing, and practice features share the same suite. That breadth makes Neiro a strong fit for people moving between those tasks; someone who needs only a narrowly focused repair tool may prefer a simpler alternative.

Key features

Separation and restoration

Neiro can split audio into vocals, instrumentals, harmonic and percussive parts, four- or six-stem mixes, and drum kits. A null-test residual accompanies each separation result, giving users an additional way to inspect what remains. Restoration covers declipping, hum removal, denoising, dereverberation, bandwidth extension, and reference mastering; noise reduction, click and crackle removal, and batch processing are also supported. This range makes it useful for varied cleanup work, though the optional neural backends involve extra downloads and license checks.

Transcription, editing, and practice

Audio can be transcribed to MIDI, with exports for MusicXML, ASCII tablature, and LRC lyrics. Studio's non-destructive waveform edits preserve a reversible workflow, while Learn combines loop regions, count-in, metronome, step mode, WebMIDI, and DAW wait mode. These additions are particularly relevant to musicians who want to move from analysis or repair into practice without leaving the suite.

DAW capture and extensions

Shared-window VST2 and CLAP injectors can capture audio into Neiro's interface, and a documented VST2 effect acts as a pass-through injector in a DAW. A local Python adapter plugin MVP offers an extension path, but granted adapters run in the Neiro process without a sandbox, so users should grant access only to adapters they trust.

Pricing

Neiro is free, with a free plan. Desktop releases do not bundle neural weights: those download on first use. The core DSP floor avoids that dependency, but users choosing a model should check its own license, since some are restricted to non-commercial or research use. The engine, desktop shell, and frontend are MIT licensed; that does not override individual model terms.

Platforms

Neiro supports Linux, macOS, Windows, web, and self-hosted use. Desktop installers are provided as Windows MSI/EXE, macOS DMG, and Linux AppImage/DEB. The Python package requires Python 3.10–3.12. WAV and FLAC work without ffmpeg, while compressed or video inputs require ffmpeg on PATH, a practical setup consideration for users working with those formats.

The browser interface binds to 127.0.0.1, and the desktop shell restricts connections to the local engine origin. The security policy says outbound network activity is off by default apart from user-initiated model downloads and updates; model downloads are checked against manifest SHA-256 values. Third-party model weights can still be dangerous. Full MUSDB18-HQ and MAESTRO evaluation numbers require users to provide the datasets, which limits what can be assessed from built-in evaluation results.

Who it's for

Neiro is best suited to musicians, producers, and audio restorers who value local processing and want separation, cleanup, transcription, waveform editing, and practice features together. It is less suitable for users who need a ready-to-use neural model without checking its terms, or who cannot meet the Python and format-specific ffmpeg requirements.

Documentation and public GitHub Discussions or Issues provide support, with private reporting available for security vulnerabilities.

Pros and cons

• Pros: Local processing keeps audio on the user's machine, and the local-only interface boundary narrows network exposure.
• Pros: Separation, restoration, transcription, editing, practice, and DAW capture give musicians a notably wide set of workflows in one suite.
• Pros: Core DSP works without model downloads, and WAV and FLAC do not require ffmpeg.
• Cons: Python 3.10–3.12 is required for the Python package, and compressed or video inputs need ffmpeg on PATH.
• Cons: Neural weights download separately and may carry non-commercial or research-only licenses; third-party weights also warrant caution.
• Cons: Python adapters run without a sandbox, and full MUSDB18-HQ and MAESTRO evaluations depend on user-provisioned datasets.

Alternatives

For a wider comparison, browse Audio Restoration Software.

• Cathar is a free option for Linux and macOS users.
• VinylRest is a free option for web, macOS, and Linux.
• Wave Corrector is a free option for Windows and Linux.
• PD Cleaner is a free option for Windows.
• CEDAR Cambridge is a paid modular hardware and software system for macOS and Windows, with custom pricing and options that vary by configuration.
• Vinyl Restoration Suite is a free option for Windows, macOS, and Linux.
• @audio/denoise is a free web option.
• Algorithmix Sound Laundry is a paid option for Windows.

Verdict

Choose Neiro if you want a free, locally processed audio workspace that spans separation, restoration, transcription, editing, and practice. Its breadth and local-first design are the main reasons to pick it; look elsewhere if you need a more focused tool, cannot manage its setup requirements, or want neural models without separate licensing checks.