Skip to main content
AI Music Generator writes and produces songs: vocals, instruments, and cover art. Describe the song you want, or write your own lyrics and style. Each generation gives you two versions to choose from. Use it for jingles, background music for videos and podcasts, birthday songs, and personal gifts.

​
Make a song

1

Open AI Music Generator

Click AI Music Generator in the sidebar, or Audio > Music Generator in the top menu. Usage at the top shows how many text-to-speech characters you’ve used this period.
2

Choose a model

V6 is the default. You can also pick V6 Wild or V6 Mini and compare; all three cost the same.
3

Describe the song

With Custom Mode off, describe it in Song Description, up to 400 characters: the genre, mood, what it’s about, and any names to include. The lyrics are written for you.For example: “An upbeat acoustic song for a yoga studio’s anniversary, about morning classes, community, and breathing easy.”Not sure where to start? Pick one under Start with a template, or click Surprise me for a random idea. Both fill in the description, and you can edit it before you generate.
4

Set the vocals

Tick Instrumental Only for music without singing. Otherwise, set Vocal Gender to Auto, Male, or Female.
5

Click Generate

Generate stays at the bottom of the form while you scroll. The song takes a few minutes. Two versions appear under Music, each with its own cover.
The Music tab with two generated versions of the song Calm acoustic folk, soft fingerpicked guitar, each with crayon-drawing cover art and Play, Share, Details, and Download buttons

​
Start from a template

Templates are ready-made sounds, from a modern pop anthem or a 90s R&B slow jam to lo-fi study beats, a birthday song, or a podcast intro. Each one has a 30-second preview made with that exact style.
1

Find a template

The most popular ones are under Start with a template. Click Browse all to open the Templates tab, where you can filter by category or search.
2

Listen first

Click ▶ on a template to hear its preview. Previews don’t use any characters.
3

Use it

Click the template. With Custom Mode on, it fills in Style, Exclude styles, and the vocals, and leaves your lyrics as they are. With Custom Mode off, it fills in the song description. Click Undo in the message that appears to go back.
If you use an instrumental template while your lyrics are filled in, your lyrics are kept and the song has vocals.

​
Write your own lyrics

Turn on Custom Mode to control the words and the sound:
  • Style: genres, instruments, mood, and tempo, up to 1,000 characters (200 on the Free plan). For example, “Calm acoustic folk, soft fingerpicked guitar, gentle female vocals, 70 BPM”. Click a tag under the field (Genre, Mood, Vocals, Instruments, Tempo, Era & sound) to add it, and click it again to remove it. Recent puts back a style you used before, and Random style picks a template’s style.
  • Exclude styles (optional): sounds you don’t want, separated by commas, such as “autotune, heavy distortion, choir”. The AI tries to avoid them, but it can’t guarantee it.
  • Lyrics: up to 3,000 characters. Mark the parts with tags such as [Verse], [Chorus], and [Bridge]: click a tag button above the field to add it where your cursor is. With the field empty, Song structure adds a full outline (intro, verses, choruses, bridge, outro) to write into.
AI Song Generator with Model Suno V6, Custom Mode on, a Style of calm acoustic folk, Lyrics with a Verse and a Chorus about rolling out a yoga mat and breathing, Instrumental Only off, Vocal Gender Female, and Generate
On the Free plan, the description is limited to 100 characters, the style and exclusions to 200 each, and the lyrics to 400.

​
Your songs

Each song has Play, Reuse prompt (pencil), Share, Details (ⓘ), and Download.
Song details: the cover, the text and lyrics, Created At, Characters Used 300, Status Completed, Generate Lyric Video, and Delete, Play, Share, and Download
  • Download offers MP3 (a smaller file, for sharing and listening) and WAV (uncompressed, for editing and mastering). The first WAV download of a song takes about a minute to prepare, and it starts on its own when it’s ready; after that it downloads straight away. WAV doesn’t use any characters.
  • Reuse prompt loads the song’s style and lyrics (or its description) back into the form, so you can change something and generate again.
  • Details shows the lyrics, the date, and the characters used, and has Delete.
  • Generate Lyric Video makes a simple video of the cover with the lyrics on screen. It’s not a music video.
  • Share makes a public link, like it does for speech. On a Teams plan you can also share with your team.
Examples shows songs other people have made; open Details on one to see its prompt, and click Use this prompt to load it into the form. All your songs are also listed in History > Music.

​
What it costs

Each generation uses 600 text-to-speech characters: 300 for each of the two versions. If a song fails, it isn’t counted. The Free plan’s 100 characters are enough for one generation a month; paid plans include many more. See Audio plans and limits.
The Usage number at the top updates when you reload the page.
You can also ask Marky to make a song in a chat.

Sound effects

Make sound effects and ambience.

Audio generator

Mix voices, music, and sound in one clip.