UVR Vocal Remover

by Anjok07AppMake video or audio

Splits a song into a vocal track and an instrumental track on your own computer

The problem it solves

You want to sing over a song, or need just its backing track for a video, and the only instrumental you can find is a muddy upload.

For Musicians and singersKaraoke and cover makersVideo editors

Before: You search for an official instrumental, or upload the song to a paid web tool and wait for the result.
After: You open the file in UVR on your computer, press Start Processing, and get separate vocal and instrumental files to use.

Stars: ★ 27k
License: MIT
Cost: Free
Platforms: Win · Mac · Linux
Runs: Local

Screenshots

From the project's own repository.

The UVR window with input and output pickers, the MDX-Net model menu, output format choices and the Start Processing button — Image: Anjok07 (MIT). Original file

Overview

UVR is a free desktop app that uses AI separation models to pull the vocals out of an audio file. You pick a song, choose a model and press Start Processing; it saves the vocals, the instrumental or both as WAV, FLAC or MP3 files. It runs on your own computer, with installers for Windows and Mac and a source install on Linux. It suits musicians, karaoke makers, video editors and anyone who needs a clean backing track or an a cappella.

Most of its separation models come from its own developers and ship inside one free app, so you need no paid web service and no coding to split a track.

Try these first

Start with the first one: it is a small task you can check easily.

Select a short song as input, tick Sample Mode (30s), keep the MDX-Net method and press Start Processing. Listen to both files.First try · on your computer
Tick Instrumental Only and save the result as MP3 to make a karaoke backing track.
Switch the process method to Demucs and split a song into separate parts such as drums, bass and vocals.
Turn on GPU Conversion and compare how long the same song takes.needs a supported GPU

Example prompts written by Agent Flows Hub.

Features

From the creator's documentation.

Feature	What it does	Runs
Vocal and instrumental split	Saves the vocals and the backing music of a song as separate audio files	Your computer
Several model families	Choose VR Architecture, MDX-Net or Demucs models, including Demucs four-stem models	Your computer
Output formats	Write results as WAV, FLAC or MP3	Your computer
Sample Mode	Process only 30 seconds to check a setting before running a whole track	Your computer
Pitch and tempo tools	Change the pitch or stretch the timing of a track	Your computer
GPU conversion	Faster processing on a supported Nvidia card or an Apple-silicon Mac	Your computer

Good to know

Separate the vocals from the backing music of an audio file you are allowed to use, for a karaoke track, a remix stem or a backing track for a video. You choose a separation method and model, and the app writes the separated parts as new audio files.

Limits

The Windows installer needs Windows 10 or later and must be installed on the C: drive, and the Mac builds need macOS Big Sur or later. On a Mac the app may not open at first; the README's fix uses Terminal commands that relax macOS app checks, so turn them back on afterwards.
Processing is heavy and its speed depends on your hardware. GPU conversion needs an Nvidia card with at least 6 GB of video memory (8 GB recommended) or an Apple-silicon Mac; AMD Radeon and Intel Arc cards use a separate DirectML version.

Notes by Agent Flows Hub (original project name: ultimatevocalremovergui (UVR)), written from the creator's documentation, not from running it ourselves. UVR (ultimatevocalremovergui): README and installation · UVR development update (issue 2399)

Stuck setting it up?

Tell us what happened. We take a look first; if it needs paid work, you get an estimate before anything starts.

Get help

How to start

Get the appDownload it from the official page, or paste the install request into an AI agent you already use.
Open itFollow the first-run steps in the app.
Give it a first taskLoad one short song, tick Sample Mode (30s) and run the MDX-Net method. Listen to both output files before processing a full track.

Steps follow the creator's instructions. The request tells your agent to read them first and wait for your OK.

Before you start

A computer with Windows, macOS or Linux.
A 64-bit computer with Windows 10 or later, or macOS Big Sur or later
Optional: an Nvidia graphics card with 6 GB or more of memory for faster processing
On Linux: Python and a few terminal commands to set it up from the source code
Free. Nothing to pay.

Details

Made by: Anjok07
License: MIT Terms
Platforms: Windows, macOS, Linux
Runs on: Your computer

View source Official docs Report a problem

For AI agents

This page as a Markdown install brief:

/en/library/uvr-vocal-remover/install.md

Other apps

Vane Runs your own AI search engine that answers questions with links to its sources
Vibe Research Gives you one local workbench to follow and research China, Hong Kong and US stocks
VoiceStudio Clones voices, dubs videos and reads text aloud on your own computer
Voyager Adds folders, a timeline and export to Gemini, ChatGPT, Claude and other AI chat sites
World Monitor Shows world news, conflicts, markets and infrastructure on one live map with AI briefs
AgentFlux Runs document workflows and AI agents on your own computer, with results saved locally

See all 113 apps

GitHub