What does the vocal remover API return?

A standard vocal remover API task returns two labeled tracks when the source supports it: vocals and no_vocals. If you set multivocal to lead_back, the task can return lead vocals, backing vocals, instrumental, and mix_no_lead. Those labels help the vocal remover API feed a preview player, export package, or timeline without guessing.

Can I use the vocal remover API for karaoke backing tracks?

Yes. A karaoke app can submit the original song, then use no_vocals as the backing track. The vocal remover API is a good fit when users expect a quick instrumental version and do not want to open a desktop audio editor. Store the instrumental with the original title so singers can replay or share it later.

Which files can I send to the vocal remover API?

The public endpoint accepts a public HTTP or HTTPS audio_url. In the playground, users can upload a local file first and the app submits the resulting URL. The vocal remover API can request mp3, wav, flac, aac, or ogg output, so choose the format your editor or download flow already supports.

How is vocal remover API pricing calculated?

Billing uses rounded-up source duration per started minute. A short clip is easier to test, while a full song costs by its rounded source length. See current per-minute rate in the pricing table on this page. Keep your own request record if your client retries a submission. The vocal remover API estimate on this page uses the same rounded-minute rule shown in the docs.

Where is the vocal remover API reference?

The full parameter list, cURL example, and response shape live at /docs/vocal-remover. Use POST /api/v1/audio/generations to create the vocal remover API job and GET /api/v1/tasks/:id to poll the result. The docs show the same model id, stem setting, and optional lead_back field used by the playground. They also show how completed output URLs and track labels are shaped for backend integration. If your app has its own queue, map the returned task id to your internal job id before storing outputs.

Vocal Remover API

Add a vocal remover API to your app. Send a public song URL, remove vocals, and return vocal plus instrumental tracks through one async task.

Input

Source audio*

Drag audio here or click to uploadMP3 / WAV / AAC / OGG · max 20 MB · up to 1 file

MP3 / WAV / AAC / OGG

Stem*

vocals

Splitter

Default auto

Extraction level

deep_extraction or clear_cut

Lead/back vocals

Optional: lead_back

Dereverb

Optional

Output format

mp3, wav, flac, aac, or ogg

Result

What you can build with this model

Real-world workflows and production use cases you can build and ship with this model.

vocals separated from an instrumental backing track

Vocal remover API for karaoke tracks

Use the vocal remover API to turn a mixed song into a backing track for karaoke apps, rehearsal tools, and creator workflows. Your product sends one audio URL and gets downloadable tracks back after the task completes. The vocal remover API keeps the user flow short: upload, wait for completion, play the instrumental.

Remove vocals

audio editor showing isolated vocal and backing tracks

Vocal remover API for remix editors

Route isolated vocals into remix, mastering, sampling, or moderation pipelines. The vocal remover API returns labeled tracks, so your editor can place the vocal and instrumental outputs in the right lanes. That label mapping matters when users export ZIP files, timelines, or preview players. Use the same labels to keep project files, download names, and moderation records consistent. The vocal remover API also helps teams keep one source task connected to every exported track.

async task queue processing vocal removal jobs

Production vocal remover API jobs

Client uploads fail, networks drop, and users refresh tabs. Keep your own task record before calling the vocal remover API, then poll the returned task id until output is ready. That keeps paid workflows and batch queues tied to one reAPI task. For batch jobs, store the source URL, selected format, and task id before showing the user a progress state.

Pricing

Credit-based — 1 credit = $0.001 USD. Pay only for completed generations.

Category	Unit	Price
Vocal removal
Source audio	1 started minute	$0.2 200 credits

Why reAPI

Why this vocal remover API is pinned to vocals

This vocal remover API page submits the stem-separator model with stem fixed to vocals. Users do not have to pick from every possible instrument when the job is simply remove vocals. The playground and docs both reflect that same constraint.

Lead and backing vocals when needed

Set multivocal to lead_back when your product needs lead and backing vocals as separate outputs. If backing vocals are not detected, the task still returns the available tracks. This is useful for rehearsal products, cover tools, and editors that need to mute only the lead vocal while keeping harmonies in context.

Vocal remover API output your app can map

Completed tasks return audio URLs and track metadata when labels are available. That makes the vocal remover API easy to wire into queues, editors, and download pages. Your backend can store the task id first and attach URLs after polling finishes.

Integrate the vocal remover API in three steps

step 01
Create an API key
Generate a reAPI key from workspace settings. The same key works across image, video, and audio endpoints, including the vocal remover API.
Open
step 02
Submit a vocal removal task
POST to /api/v1/audio/generations with model audio-stem-separator, stem vocals, and a public audio_url. Add output format or lead_back only when your workflow needs it.
Open
step 03
Poll and download
GET /api/v1/tasks/:id until the task is completed. Download output.audio_urls or read output.tracks to map vocals, no_vocals, and optional lead/back labels.
Open

View all models

start building

Ready to ship?

Try it in the playground or grab an API key to integrate now.

Try vocal removal View vocal remover docs

What you can build with this model

Real-world workflows and production use cases you can build and ship with this model.