Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

NEUTRINO

studio-neutrino.com

  • Free plan
  • Windows
  • macOS
  • Linux
  • Web

What NEUTRINO is

NEUTRINO is a free AI singing voice generator that turns lyrics and melody into synthesized vocals. Its AI estimates vocal timing, pitch, voice quality and breathiness from sheet music, then a vocoder produces the singing. The project says its models learn vocal traits from recordings made with performers’ permission or audio released by public institutions, rather than from existing pretrained models. The FAQ describes Muon voice retakes and adjustable quality and processing speed, with advanced, standard and elements modes. Advanced and standard outputs are 48 kHz, 16-bit mono WAV; elements output is 24 kHz, 16-bit mono WAV. NEUTRINO is listed for Linux, macOS, Windows and web. Users download the software and add a voice library. It is free for general music and audio production, but embedding it or using generated audio inside an application requires separate permission. Character-specific terms also apply to ordinary audio-file uses and vary by character and asset type. The FAQ says users cannot currently create their own voice libraries with a released framework.

Who it suits

NEUTRINO suits music and audio producers who want to generate singing from lyrics and melody, and can use its supplied voice libraries. Application developers should note that embedding or in-app use requires separate permission.

Good to know, and what to watch

  • Free for general music and audio production
  • Supports adjustable quality and processing speed
  • Offers voice retakes in Muon
  • Advanced and standard output is 48 kHz mono WAV
  • Embedding requires separate permission
  • In-app use of generated audio requires separate permission
  • Users cannot currently create their own voice libraries

NEUTRINO

Free

NEUTRINO freeware

Free

Free plan

Getting started

  1. Open the NEUTRINO website
  2. Download NEUTRINO
  3. Install the software on Linux, macOS, Windows, or use the web version
  4. Add a voice library
  5. Enter lyrics and melody or provide sheet music
  6. Choose Muon quality and processing settings
  7. Generate and export a WAV file

Limits to check

The free plan covers general music and audio production. Embedding NEUTRINO or using generated audio inside an application requires separate permission, and creating personal voice libraries is currently unavailable through a released framework.

The Club's review

NEUTRINO is a sheet-music-driven AI singing synthesizer for composers, producers and audio creators who want a sung part without recording a vocalist. It is strongest when your source is written lyrics and melody, and less suitable when you need vocal conversion, custom voice cloning or software-embedded generation.

The free price removes a major barrier for ordinary music production. The important qualification is licensing: application integration and the use of particular character voices or assets require permission and character-specific terms.

Overview

Enter lyrics and a melody, then NEUTRINO estimates timing, pitch, voice quality and breathiness from the notation before a vocoder produces the vocal. This notation-first approach suits songwriting, arranging and demonstrations where the musical part exists on paper or in a score. It does not accept vocal input and does not provide voice cloning, so it is not a replacement for a vocal-conversion service.

Setup involves downloading NEUTRINO and adding a voice library. The project says its models are trained from recordings made with performers’ permission or audio released by public institutions, rather than from existing pretrained models. That provenance is relevant to creators who need a clearly described training approach, although each voice library and character still carries its own terms.

NEUTRINO itself permits ordinary audio-file use for music and audio production without a restriction from the software. Embedding the software, or putting generated audio inside an application, needs separate permission. Character-specific conditions also apply and can differ by character and asset type, so a commercial release should be checked voice by voice.

Key features

Generation controls

Muon provides voice retakes plus adjustable quality and processing speed. Its advanced, standard and elements modes let you choose a processing approach rather than forcing one fixed render. Advanced and standard exports are 48 kHz, 16-bit mono WAV; elements exports are 24 kHz, 16-bit mono WAV. Choose the former when the stated 48 kHz output matters to your production, and elements when its lower-resolution render is sufficient.

Voice character

The project says its system extracts traits from vocalist recordings and reproduces voice quality, singing style and vocal quirks. That can give a written performance more character than a generic synthetic tone, but it is not user-controlled cloning: you cannot create your own voice library through the released framework. The project cites quality-control and support-cost concerns for that limitation.

Hardware and distribution

About 4–6 GB of GPU VRAM is described as sufficient; processing may run out of memory below 4 GB. CPU processing modes are also described, giving users without a suitable GPU another route, though potentially with a different speed trade-off. The FAQ names JOYSOUND Uta-Suki Music Post, BIG UP!, TuneCore Japan and Dwango’s music monetization service as supported distribution services as of May 2025.

Pricing

NEUTRINO: free. The free plan covers general music and audio production, making it a sensible starting point for composers, hobbyists, educators and producers who can work with the supplied voice libraries. There are no paid tiers or subscription renewal to weigh against it.

The free price does not grant blanket product-embedding rights. Integrating NEUTRINO into software or using generated audio inside an application requires separate permission, and character-specific terms govern particular voices and asset types. Those permissions are the main commercial consideration, not a monthly quota. Users needing a custom voice library also cannot build one with the released framework.

Platforms

NEUTRINO is offered for Linux, macOS, Windows and the web. The desktop options suit score-based production on a local machine, while web access may be preferable for users who want a browser workflow. Output is WAV, with the mode-specific sample rates and bit depths described above.

Who it's for

Choose NEUTRINO if you write melodies and lyrics, need a free synthesized singer for music or audio files, and can work within the available voice libraries. It is particularly well matched to composers prototyping arrangements and producers who want controllable retakes without hiring a vocalist for every draft.

Look elsewhere if your starting point is a recorded vocal, if you need to clone your own voice, or if your product must generate vocals inside an application under a single, uncomplicated licence. Low-memory systems should also check the stated VRAM guidance before committing to GPU rendering.

Pros and cons

Pros

  • Free for ordinary production: music and audio-file use has no restriction from NEUTRINO itself.
  • Notation-based workflow: lyrics and melody are enough to create a sung part without vocal recording.
  • Useful render choices: retakes and quality/speed controls support different production priorities.
  • Clear WAV outputs: advanced and standard modes provide 48 kHz, 16-bit mono files.
  • Broad platform coverage: Linux, macOS, Windows and web access cover common working environments.

Cons

  • No vocal input or cloning: it cannot transform a supplied vocal or make a personal voice library.
  • Integration requires permission: application developers need approval beyond the free production licence.
  • Character terms vary: commercial use depends on the specific voice and asset type.
  • Hardware can matter: less than 4 GB of GPU VRAM may cause memory failures.
  • Library creation is unavailable: the released framework does not currently let users publish their own voices.

Alternatives

Revocalize AI is the better fit when you want conversion minutes, a starter AI voice and a freemium workflow spanning mobile, web, desktop, API and self-hosted options; its free plan allows five conversion minutes per month.

SightSinger suits a browser-only workflow and offers a paid Solo plan at $6 a month for users who prefer a subscription service.

Vocalize is the choice for API-led production, charging $0.048 per generated minute plus $0.003 per request with private voice clones and no subscription.

Kits AI is preferable when you need generative vocals and voice-design tools across web, Windows and API; its free plan includes 15 conversion minutes but no voice slots or download minutes.

SoulX-Singer is an alternative for researchers and developers who want free code and model weights under the Apache 2.0 licence.

VoiSona Song is worth considering if you want a freemium singing editor for iOS, macOS or Windows.

Audimee fits browser users seeking royalty-free voices and MP3 downloads; its free plan is limited to personal use, 15 conversion minutes and no custom voice slots.

Tunela is the better match for a broader browser toolkit combining TTS, stem separation and video features with small free quotas.

For a wider comparison, see AI Singing Voice Generators.

Verdict

NEUTRINO is a strong choice for composers and producers who want a genuinely free, notation-based singing voice for ordinary audio production, with useful render controls and clearly defined WAV outputs. Look elsewhere for vocal conversion, personal voice cloning or an application product whose licensing must be simple and uniform; those needs run into NEUTRINO’s missing input and cloning features and its separate integration and character permissions.

The verdict

NEUTRINO offers a free route to synthesized vocals for ordinary music and audio production. Review the separate permissions and character-specific terms before using it in an application or with particular voice assets.

Questions people ask

Is NEUTRINO free?

Yes. The NEUTRINO plan costs nothing and is free for general music and audio production.

Which platforms support NEUTRINO?

NEUTRINO is available on Linux, macOS, Windows, and the web.

What files does it export?

Advanced and standard modes export 48 kHz, 16-bit mono WAV. Elements mode exports 24 kHz, 16-bit mono WAV.

Can I use generated audio commercially?

NEUTRINO places no restriction on ordinary audio-file uses such as music production, but character-specific terms also apply and vary by character and asset type.

Can I embed NEUTRINO in an application?

Embedding the software or using generated audio inside an application requires separate permission.

Does it support voice cloning?

The listed voice-cloning capability is no; users also cannot currently create their own voice libraries with a released framework.

What the maker publishes

AI-training restriction
The EULA forbids using generated audio, intermediate data or voice libraries as training or fine-tuning data for other AI singing, speech or voice-conversion systems. source
API and SDK
The official FAQ and download materials describe command-line scripts, MusicXML input and Orbit, but do not publish a developer API or SDK. source
Audio output
The standard output is 48 kHz, 16-bit mono WAV, with an elements mode at 24 kHz. source
Commercial use
The FAQ says NEUTRINO itself places no restriction on ordinary audio-file uses such as music production, but character-specific terms also apply and vary by character and asset type. source
Commercial use
Generated audio may be used commercially or non-commercially, subject to each voice library's separate terms. source
Data provenance
The maker says training uses performer-authorized recordings or public institutional datasets and is trained from scratch without pre-trained models. source
Desktop and online support
The maker provides Windows, macOS, Linux and Google Colab distributions. source
Developer and release
Independent reference material identifies SHACHI as developer and dates the first public release to February 22, 2020. source
Documentation and support
The official FAQ covers installation, rendering errors, GPU issues, file formats and licensing, and directs unresolved questions to a contact form. source
Download sizes
The current download page lists roughly 3.05 GB for Windows Tau, 1.76 GB for macOS Tau, 1.70 GB for Google Colab Tau and about 346–349 MB for individual Tau voice libraries. source
Free availability
The software and many voice libraries are free to download and use for ordinary music and audio production. source
Free use
NEUTRINO is free to download and use for general music and audio production, while embedding the software or using generated audio within an application requires separate permission. source
Generation history
The product has progressed from Electron to Muon and then Tau generations, with Muon introducing diffusion models and Tau adding further stability and expression improvements. source
Generation options
The FAQ says Muon supports voice retakes and adjustable quality and processing speed, with advanced, standard, and elements modes. source
GPU fallback
Older unsupported NVIDIA GPUs can be bypassed by running the renderer in CPU mode. source
GPU memory
The FAQ says 4–6 GB of GDDR VRAM generally works, while less than 4 GB can cause memory failures and 6 GB or more is usually sufficient. source
Hardware requirements
The FAQ says about 4–6 GB of GPU VRAM is sufficient, while some processing may run out of memory below 4 GB; it also describes CPU processing modes. source
How it works
Its AI estimates vocal timing, pitch, voice quality, and breathiness from sheet music, then a vocoder synthesizes the voice. source
Input and workflow
The project says users can make it sing by entering lyrics and melody, and its download page instructs users to download NEUTRINO and add a voice library. source
Input format
MusicXML is the primary score format; the FAQ suggests converting MIDI or other project formats through tools such as UtaFormatix. source
Integrations
The FAQ lists JOYSOUND Uta-Suki Music Post, BIG UP!, TuneCore Japan, and Dwango’s music monetization service as supported distribution services as of May 2025. source
Licence status
NEUTRINO is freeware under a maker-owned EULA; the EULA assigns copyright to SHACHI and prohibits unapproved redistribution and reverse engineering. source
Limit: DAW integration
SoundBounty says NEUTRINO is not a DAW plug-in and requires exporting WAV audio for DAW use. source
Limit: language coverage
SoundBounty reports that lyrics are entered in Japanese hiragana or katakana and that English and Chinese are not currently supported. source

Checked against the maker's pages on 21 Oct 2026.