Questions, answered in the chat.
What people ask first, each answer with the chat that does it.
Practise with any song
How to pull out the vocals for a practice track
Bring the song to the chat and say “take the vocals out so I can practise”. audioworker splits it into vocals, drums, bass and other, mutes the vocals, and the band plays on without the singer.
How to make a karaoke version of any song
Ask the chat to remove the vocals, listen in the player, then ask for the download. The karaoke version is the instrumental exactly as you hear it, as an MP3 or a WAV.
How to get an a cappella from a song
Split the song and ask for “only the voice”. The vocals play on their own, and the vocal stem downloads as a WAV or an MP3 for a remix, a mash-up or practice.
How to remove the drums from a song to play along
Split the song and say “mute the drums”. The rest of the band plays without them, so you can play your own part, with a click and a slower tempo if you need one.
How to isolate the bass line of a song
Split the song and solo the bass. You hear the bass line on its own, and you can slow it down, loop a bar, or ask for it as MIDI to see the notes.
How to change the key of a song to fit my voice
Ask for the key up or down in semitones, for example “down two semitones”. The pitch moves, the tempo stays the same, and the download keeps the new key.
How to slow down a song without changing the pitch
Use the Speed menu in the player, from half speed (0.5×) up to 1.2×. The song plays slower or faster at the same pitch, so you learn the real notes.
How to loop the hard part of a song
Say what to loop, like “loop the chorus” or “loop 1:20 to 1:35”. The player repeats that part until you stop it, at any speed and key.
Split, read and download
How to split a song from a YouTube link
Paste the link into the chat and say what you want. audioworker fetches the audio, splits it into four stems, and loads them in the player.
Which audio and video files can I split?
Most audio and video files work: MP3, WAV, M4A, AAC, FLAC, OGG, Opus, WMA and AIFF, plus MP4, M4V, WebM, MKV, MOV and AVI, up to 400 MB and 90 minutes.
How to find the key, tempo and chords of a song
Ask “what key is this in, and what are the chords?”. The chat reads the key, the tempo and the chords from the song, and the chords appear along the waveform.
How to add a click track to a song
Ask for “a click track”. It arrives as its own stem called Click, following the song’s beat, and stays muted until you press its M button.
How to download one stem as a WAV file
Ask for “the bass stem as a WAV” (or vocals, drums, other). The file appears in a Downloads card in the chat, ready to save.
How to download just one section of a song
Set the section (“the chorus” or “1:00 to 1:30”) and ask for the download. You get that section only, mixed exactly as you hear it.
How to turn a part of a song into MIDI
Ask for “the bass line as MIDI”, or any other stem, or the whole mix. You get a MIDI file of the notes it hears, to open in your music app.
Make songs with AI
How to make a song with AI from a description
Type what the song is about and how it should sound. The AI writes the lyrics (or uses yours), picks a voice, and makes versions to listen to, from 10 seconds to five minutes long.
How to make an AI song with my own lyrics
Paste your lyrics into the chat and say the style and the voice. Tick the box that says the words are yours, and the AI sings them as written.
How to make an AI song in another language
Write your request in the language you want, or name it: “in Korean”. Songs can be sung in English, Korean, Japanese, Chinese, Spanish, French, German, Portuguese and Italian.
How to make an instrumental song with AI
Say “no vocals” or “instrumental” with the style and the length. You get music with no singing, from 10 seconds to five minutes.
How to turn a photo into a song
Drop one to ten pictures into one message, add a line if you want (“a summer song for the car”), and audioworker turns them into a song.
How to make a song that sounds like a reference track
Give the chat a reference (a file, a YouTube link or a track in your library) and say what the new song is about. The reference shapes the sound and the mix; it is never a cover.
How to turn a hummed melody into a song
Say “make a song from my hum”, pick a speed and hum your tune after four clicks. Answer one question (words the AI writes, or your own) and the AI makes a whole song, with a singer on the notes you hummed.
How to change one line of lyrics in an AI song
Quote the line and the new words: “change ‘sky is blue’ to ‘sky is grey’”. Only that line is sung again, on the same melody, and the rest of the song stays as it was.
Add voices and parts
How to add AI vocals to an instrumental
Ask for vocals on your track: “a female voice singing about leaving town”. The AI sings over the music you have, and the take arrives as a new stem in the player.
How to add a backing choir to the chorus
Ask for “a mixed choir on the choruses”. The AI finds the choruses, or uses the times you give, and adds a backing choir on up to four sections as one new stem.
How to switch a male vocal to a female voice
Say “switch this vocal to a female voice” (or male). The same words and melody are sung again by a stock AI voice, and the result arrives as a new stem.
How to sing a song in my own voice
Record your voice once in the chat (one take of 10 to 30 seconds) and tick the consent boxes. Then say “sing it in my voice” on any vocal, and it comes back as “My version”.
How to add a new instrument part to a song
Ask for the instrument and the style: “add warm strings”. The chat asks where (where it fits best, the whole song, or your times) and the new part arrives as its own stem.
What happens when the AI makes two versions?
You listen to each version, then keep one or keep all. A kept song becomes a track with stems; a kept AI part stays in your track and the others drop out.
Mix, edit and finish
How to mix a song by chatting
Say what you want to hear, like “vocals up a little, drums down”. The player changes the mix while the song plays, and every change is saved with the track.
How to add reverb to only the chorus
Say “reverb on the vocals in the chorus only”. The chat finds the chorus, puts the effect on that range of the vocal stem, and you hear it at once.
How to use the EQ to fix a muddy mix
Say what sounds wrong: “it sounds muddy” or “too harsh”. The chat sets the 10-band EQ on the mix, such as a cut around 250 Hz for mud, and you hear it right away.
How to cut or fade out part of a song
Say the edit: “cut the intro”, “fade the vocals out at the end” or “mute the drums for the first verse”. It plays at once, can be undone, and rides every download.
How to make a song loud enough for Spotify
Ask for “a file ready for streaming”. You get your mix, as you hear it, at −14 LUFS, the loudness streaming services play at, with the measured numbers.
How to master a song online
Ask the chat to “master it”. On Studio, it masters your track, or makes an HD master of the mix as you hear it, with your levels, mutes and edits.
How to clean up a noisy or echoey vocal
Split the track and ask “clean up the noise on the vocals” or “take the room echo out”. On Studio, the clean-up runs on that stem and the result comes back to the player.
How to save a mix and come back to it
Say “save this mix as Live version”. Your levels and settings are kept by name, and every track and mix is in your library, from any chat.
Plans, rights and privacy
Is audioworker free?
Yes. The Free plan splits 20 songs a day, with mixing, effects, key and tempo, loudness-matched files and downloads. AI songs and the other AI tools are on Studio.
What do I get with Studio?
Studio is $9.99 a month. It unlocks every AI and HD tool with 60 HD and AI minutes a month, 500 minutes of splitting, 100 YouTube links, 150 messages a day and 50 GB, kept until you delete it.
How long are my songs kept?
On Free, tracks are kept for 24 hours after your last activity. On Studio, they are kept until you delete them, with 50 GB of storage.
Who owns the songs I make with AI?
You do. You keep your rights in what you upload and in what you make, AI-generated audio included, and audioworker claims no ownership.
Is my music used to train AI?
No. Your audio, pictures, chats and everything you make are never used to train AI models, and your tracks stay private to your account.
Do I need the rights to a song to split it?
Yes. Before anything is processed, a box asks you to confirm you have the rights to use the song, such as your own recording or a track licensed to you.
Which languages does audioworker speak?
You can chat in any language you write in. Songs and vocals can be sung in English, Korean, Japanese, Chinese, Spanish, French, German, Portuguese and Italian.