# AI vocal remover and acapella extractor

Pull clean vocals and an instrumental out of any song or video.

## How Mel-Band RoFormer isolates singing without phase swirl

Audo's AI vocal remover and acapella extractor separates songs and videos (WAV, MP3, M4A, AAC, FLAC, OGG, Opus, WebM, AIFF, MP4, MOV, MKV, WebM, AVI) up to 10 GB without an account, with the first 45 seconds free 3 times a day and no sign-up. Older vocal removers cancelled the center channel, which wiped out kick drums, basslines, and snare hits along with the singer while leaving stereo vocal reverb behind.

Instead of center-channel cancellation, Audo runs a Mel-Band RoFormer transformer across 60 overlapping mel frequency bands at 44.1 kHz stereo. Because human singing has harmonic overtones and vibrato patterns that differ from cymbals and synths, the attention layers track the vocalist even when guitars and keys play in the same frequency range.

Once the vocal waveform is predicted, Audo computes the instrumental by sample-accurate subtraction from the original stereo file. That means the backing track retains every transient and stereo detail of the master recording, ready for remixing in Mix audio or cutting in the Audio editor.

Tool: Split stems (`split_audio`). The same tool's other pages: https://audo.ai/stem-splitter.md, https://audo.ai/karaoke-maker.md, https://audo.ai/remove-music-from-video.md. Price: 8 credits a minute — $0.08/min ($4.80/hr), with no subscription — of input for whole files, charged only when a job succeeds.

## Free API

```bash
curl -F file=@song.mp3 -F stems=vocals https://audo.ai/api/try/split-stems -D response.headers -o response.body
```

Read response.headers and response.body before naming media. Handle 202, 200 JSON and errors with the complete response client at https://audo.ai/docs/api?utm_source=markdown&utm_medium=agent#responses.

Free for the first 45 seconds of each file, 3 times a day per person, with no key or account. Uploads up to 10 GB, 2 at once from one address. The response is the result itself; a job that runs longer than about 30 seconds returns 202 with a link to check on it. With an API key (`-H "Authorization: Bearer $AUDO_API_KEY"`), omitting free runs the whole file with credits if the balance covers it, otherwise the free portion if quota remains. Set free=true to prohibit spending; for paid whole files use free=false with an approved positive max_credits. A free result cut short comes with finishUrl (and the Audo-Finish-Url header when the response is the file itself): give it to the person, who pays there (from $5, no sign-up), and the whole file runs; keep checking the status link, whose finishedBy names the whole-file job and its own status link. Details: https://audo.ai/docs/api.md

## Options

Form fields with the same names as the MCP tool's inputs:

- `stems`: Which stems to separate: vocals (vocals first, then instrumental), instrumental (instrumental backing track first, then vocals), four_stems (vocals, drums, bass, and other), or dialogue (music-removed mix first, then dialogue, music, and sound effects). Default vocals.
- `output_format`: The result's format: same (as the input; a video stays a video), wav, mp3, m4a, or flac. Default same.

## Languages

Any language: it works on the sound, not the words.

## From an assistant (MCP)

Add https://audo.ai/mcp and sign in (setup for each assistant: https://audo.ai/setup.md). The tool is `split_audio`:

> Separates a song or video into true stereo stems: vocals and instrumental (vocals or instrumental first), 4 stems (vocals, drums, bass, and other), or dialogue, background music, and sound effects (with a music-removed mix first). Works on video: when output_format is same, the first stem keeps the original video picture. Set free to true to do the first 45 seconds at no cost, up to 3 times a day; otherwise it costs 8 credits a minute, charged only if the job succeeds. Returns job_id; then call get_job.

Example: "Use Audo to isolate the vocals and instrumental from track.flac."

## More

- Every tool: https://audo.ai/llms.txt
- OpenAPI: https://audo.ai/api/openapi.json
- This page for people: https://audo.ai/vocal-remover?utm_source=markdown&utm_medium=agent