Audio separation for music

Separate vocals from music, from one master to the whole catalog

A clean acapella and a sync-ready instrumental from the finished mix. No session files needed.

Now playingPop
0:00
Vocals
0 dB
Instrumental
0 dB
Drag a record onto the player
  • Pop
  • R&B
  • Hip-hop
  • Country

  • Batch

    One song or the whole catalog, in Studio or by API

  • Live

    Concerts and broadcasts in real time. Enterprise, on request

Built for

From the studio to the catalog

Sync

Instrumentals and vocal‑up versions from the original master.

Remixes

Acapellas for remixes and DJ packs.

Karaoke

Sing-along tracks at catalog scale.

Catalogs

Missing versions for legacy releases.

Live

Concerts and broadcasts, separated in real time, for enterprise teams.

The problem

Your catalog has the mix. Not always the stems.

Older releases, acquired catalogs and finished masters often come without separate vocal and instrumental files. Rebuilding them by hand takes studio time you don't have for every request.

CAMB.AI is a vocals and instrumental separator: it uses vocal isolation to give you both versions back from the mix you already own.

Live · Enterprise

Separate vocals live, on stage and on air

Connect a concert or broadcast feed and get the vocals and the band back as two separate streams, in real time. We set it up with your team, on your own feed.

  • Voice and background as two live streams
  • Any language, several speakers at once
  • Straight into CAMB.AI live translation and dubbing
Concert feedMove the level to bring the band back0:00
Vocals
Band

Quality

Measured, not just promised

In our benchmark against AudioShake, ElevenLabs and LALAL.AI, CAMB.AI kept the most voice in the foreground stem and left the least behind in the background.

The least voice left behind

14% less than AudioShake, 16% less than ElevenLabs, 29% less than LALAL.AI: a background stem you can really reuse.

Average probability that speech is still heard in the background stem.

CAMB.AI0.376%
AudioShake0.437%
ElevenLabs0.447%
LALAL.AI0.529%

CAMB.AI2.84 dB
LALAL.AI2.81 dB
ElevenLabs1.64 dB
AudioShake1.01 dB

The cleanest voice of the four

2.8× AudioShake and 1.7× ElevenLabs on how cleanly the voice is pulled from the mix.

SI-SDR, estimated by SQUIM: how much of the separated voice is the voice and not what was around it.

CAMB.AI0.376%
AudioShake0.437%
ElevenLabs0.447%
LALAL.AI0.529%

The least voice left behind

14% less than AudioShake, 16% less than ElevenLabs, 29% less than LALAL.AI: a background stem you can really reuse.

Average probability that speech is still heard in the background stem.

CAMB.AI1.895
LALAL.AI1.852
ElevenLabs1.544
AudioShake1.527

The voice that sounds best

24% above AudioShake and 23% above ElevenLabs on how good the separated voice sounds.

PESQ, estimated by SQUIM: perceived speech quality, without a reference recording.

CAMB.AI0.744
LALAL.AI0.740
ElevenLabs0.726
AudioShake0.697

The voice that is easiest to understand

Every word stays clear, ahead of all three.

STOI, estimated by SQUIM: how intelligible the separated speech is.

CAMB.AI0.258
LALAL.AI0.255
ElevenLabs0.229
AudioShake0.209

The most voice in the voice stem

23% above AudioShake and 13% above ElevenLabs.

Average CLAP speech-focus gap: how clearly the foreground stem is about speech.

MeasureCAMB.AIAudioShakeElevenLabsLALAL.AI
Separation quality (SI-SDR)2.84 dB1.01 dB1.64 dB2.81 dB
Voice left in the background (Leakage)0.376%0.437%0.447%0.529%
Perceptual quality (PESQ)1.8951.5271.5441.852
Intelligibility (STOI)0.7440.6970.7260.740
Voice kept in the foreground (CLAP gap)0.2580.2090.2290.255
DNSMOS speech quality (SIG)2.6622.3872.5232.684
DNSMOS background quality (BAK)2.8842.3072.5382.973
DNSMOS overall quality (OVRL)2.3031.9232.0942.327

Internal benchmark: CAMB.AI, AudioShake, ElevenLabs and LALAL.AI on the same seven clips from Wikimedia Commons (gaming commentary, a film trailer, live music, an outdoor interview, a live newsroom, archival TV and a red carpet interview), averaged. SQUIM and DNSMOS are reference-free estimates. On the three DNSMOS scores LALAL.AI is slightly ahead; CAMB.AI leads every other measure.

Use cases

Built for how music teams work

Instrumentals and vocal‑up versions for sync

Instrumentals and vocal‑up versions for sync

Answer a sync brief fast with an instrumental made from the original master, and the vocals to build a vocal‑up mix. An instrumental maker for trailers, ads, TV and games.

Music videos and documentaries

Music videos and documentaries

Separate dialogue and singing from the soundtrack to subtitle or dub music films and documentaries into new languages.

AI dubbing
Acapella extractor for remixes and DJ packs

Acapella extractor for remixes and DJ packs

Clean acapellas for remix campaigns, DJ packs and producer contests, without asking the studio for the session files.

Karaoke maker for whole catalogs

Karaoke maker for whole catalogs

Turn a catalog into karaoke and sing-along tracks, with the original instrumental.

Catalog backfill

Catalog backfill

Create missing vocal and instrumental versions for legacy releases, in bulk.

What you get

AI stem splitter: vocals and instrumental from every song

Vocals

The vocals on their own, without the music.

Instrumental

The full arrangement, without the vocals. Drums, bass and other instruments stay together in the instrumental.

Vocal separation API

From one song to the whole catalog: run separation one file at a time in Studio, or send your catalog through the API and get results back in bulk.

POST /apis/audio-separationGET  /apis/audio-separation/{task_id}GET  /apis/audio-separation-result/{run_id}POST /apis/audio-separation-results
See the audio separation API

Live, on stage and on airEnterprise

Separate the vocals from the band in real time, for concerts, live broadcasts and streams. An enterprise feature, set up with your team on request.

Talk to us

Your masters stay yours

CAMB.AI is SOC 2 compliant and doesn't train its models on your audio. Your runs are listed in your History and you can delete them whenever you want. Separation doesn't change who owns the music or the rights to use it.

More audio separation

How it works

Extract vocals from a song in three steps

01

Upload

Add a song or a whole folder: .wav, .flac, .mp3, .aac, or video as .mp4 or .mov.

02

Separate

CAMB.AI splits each file into vocals and instrumental.

03

Download

Get both stems as FLAC, WAV, MP3 or AAC. FLAC by default, for lossless quality.

Pricing

Pick your hours. Both stems in every plan.

Every plan gives you the voice and the background as two stems, in Studio and by API. Try any plan free for 3 days, with 30 minutes of audio.

10 h
Billing

Solo

$40/month

$480 billed yearly

3 hours of audio a month

For creators and hobbyists. Pull the voice out of songs, podcasts and videos in minutes.

Start free trial
  1. Today 30 minutes to try. $0 today.
  2. Day 3 Your plan starts: $480 a year.

Session

Recommended

$130/month

$1,560 billed yearly

10 hours of audio a month

For producers and content creators working every week. Enough for full albums, podcasts and soundtracks.

Start free trial
  1. Today 30 minutes to try. $0 today.
  2. Day 3 Your plan starts: $1,560 a year.

Catalog

$375/month

$4,500 billed yearly

30 hours of audio a month

For professionals and teams at volume. Clean voice and background stems for music, post-production and localization.

Start free trial
  1. Today 30 minutes to try. $0 today.
  2. Day 3 Your plan starts: $4,500 a year.

Enterprise

High volumes, live separation on your own feeds, or separation inside your own workflow.

  • Live separation, in real time
  • Volume pricing
  • SLA and dedicated support
  • Custom delivery and storage

FAQ

Questions about vocal separation

Upload the song in Studio or send it through the API. CAMB.AI returns two stems: the vocals and the instrumental.

Yes. Every run gives you both stems.

No. CAMB.AI is an AI stem splitter for two stems: vocals and instrumental. Drums, bass and every other instrument stay together in the instrumental stem.

Upload the song and download the instrumental stem. You can use it for sync, karaoke or as a backing track.

Upload the song and download the instrumental: it keeps the original arrangement without the vocals. For a whole catalog, send the songs through the API.

Upload the song and download the vocal stem: the voice on its own, without the music.

Usually an instrumental of the original master, and often the vocals to make a vocal-up or acapella version. CAMB.AI gives you both from the finished master.

Upload .wav, .flac, .mp3, .aac, .mp4 or .mov. Download FLAC, WAV, MP3 or AAC.

Yes. Download both stems as WAV or FLAC and drop them into any DAW, like Pro Tools, Logic or Ableton.

22.6 seconds to process a 30-second video. Files are processed in the background, so you can send many at once.

Yes. Use the API to send files in bulk and collect the results in bulk.

Yes. Create a separation task, check its status and get both stems back, one song at a time or in bulk.

Yes, as an enterprise feature. Talk to us and we set it up on your feed: you get the vocals and the music as two separate streams in real time.

No. CAMB.AI doesn't train its models on your audio.

Separating audio doesn't give you any rights to it. Make sure you have the rights you need for how you plan to use the stems.

Yes, as long as you hold the rights to the recording for that use.

Bring a song. Leave with two stems.

  • SOC 2 compliant
  • Your audio isn't used to train our models
  • Batch and live