What do you hear?

A mood, a melody, a sonic landscape — let's bring it to life.

Let AI guide your music creation.
Templates

AI Song Generator

Turn one idea or your own lyrics into a complete song with vocals, verses, choruses, and instrumentation. Then remix, extend, split stems, and master it on one canvas.
How it works

How to Make a Song with AI in 3 Steps

1Step 1

Generate

Chat with the Music Agent or fill in Custom mode — describe a mood, story, or style, or paste your own lyrics. You get complete songs with vocals, and every idea can branch into directions you compare side by side.

2Step 2

Shape

Treat the first take as a draft. Remix the style, extend the arrangement, rewrite one section on the waveform, split stems, regenerate the cover — every step stays connected on the canvas.

3Step 3

Release

Master the final mix, download your song cleared for commercial use, and share a public track page others can remix.

The workflow

One AI Song Generator, from Idea to Release

Describe a song or bring your own lyrics and get a complete track with vocals. Then keep going — shape it, split stems, master it, and release it without leaving the canvas.

Chat with an agent, or take full control

In Agent mode, describe the song — or attach a photo, a reference track, or a MIDI sketch — and the Music Agent drafts style directions and writes lyrics when you need them. Switch to Custom to set every field yourself: style, lyrics, title, vocals or instrumental, voice, and model.

The right model for every song

Aether covers the widest range of styles and adds style weight, weirdness, and excluded-styles controls. Reverie sings your lyrics exactly as written, across the most vocal languages. Echo builds tightly structured full songs with precise tempo. Switch models on any generation.

A canvas, not a list

Every session is a visual canvas: style directions (Music DNA), generations, edits, and stems all connected as a family tree. Branch into new directions whenever you like, generate them side by side, and keep what works.

Extend a song, rewrite a section

Remix a track into a new style, extend it from the point you choose, or select a few seconds on the waveform and rewrite just those lyrics — the rest of the song stays as it was, and the lyrics stay time-aligned. Stuck on a line? The AI assistant can inspire, polish, or extend it.

Split any track into stems

Separate vocals, drums, bass, and instruments into 2, 4, or 6 stems with one click. Each stem lands on the canvas with its own mute, solo, and volume, so you can render a custom mix — and when you need stem files for your DAW, the Stem Splitter exports them.

Master where you make

AI mastering on the same canvas: pick one or more reference sounds, compare each master with the original, and export WAV (16- or 24-bit) or MP3 at release loudness. Mastering a track you made elsewhere? AI Mastering matches it to your own reference track.

See inside your music

Karaoke-style synced lyrics, a live spectrum, and a full analysis panel — key, BPM, chord progression, loudness — for every track you generate.

Own what you make

Download your songs cleared for commercial use on any paid plan — Pro adds a license certificate in your name. Regenerate the album cover until it fits, and share a public page others can remix.
Who it's for

An AI Song Maker for Producers, Songwriters & Creators

From producers sketching ideas to artists shipping releases — the workflow adapts to where music fits in your life.

Producer comparing AI-generated style directions on a track canvas
01

Music producers

Beat block, solved: generate reference tracks, split them into stems to study or sample, and audition four style directions before you ever open your DAW.

Songwriter turning written lyrics into a sung demo
02

Songwriters

Hear your lyrics sung back in minutes. Paste your words, pick a voice and style, then iterate on melody and arrangement until it matches the song in your head.

Content creator scoring a video with licensed AI music
03

Content creators

Licensed, royalty-free music for videos, streams, and podcasts — with stems, so the vocals can duck under your voiceover and the beat can carry the outro.

Independent artist preparing a mastered, licensed release
04

Independent artists

Go from idea to release-ready without a studio budget: master the final mix, grab your commercial license and cover art, and publish on your terms.

Keep creating

More Ways to Make a Song with AI

FAQ

AI Song Generator FAQ

Describe the song you want — or paste your own lyrics — and pick a direction. Lacuna generates a full song with vocals and instrumentation, then keeps the whole journey on one canvas: remix, extend, rewrite sections, split stems, and master until it is release-ready.

Yes. Sign in and you get free monthly credits to generate and listen to full songs with vocals. Downloads and commercial use come with any subscription plan, and Pro adds a license certificate in your name.

Yes. Generate songs with AI vocals, lyrics, melody, and instrumentation, or switch to instrumental when you want music without a singer. Vocal language support depends on the model you choose.

Yes. Leave the lyrics out and the Music Agent writes them, or open the lyrics panel and use AI Muse to inspire, polish, or extend your lines. Want to work on the words first? Try the AI Lyrics Generator.

Any genre or blend you can describe — pop, rock, hip hop, R&B, K-pop, EDM, lo-fi, folk, cinematic, and more. Vocal languages depend on the model: Reverie covers the most, with 10+ languages, and every model sings in English, Spanish, Portuguese, French, Japanese, and Korean.

No. Describe the song in everyday words — a feeling, a story, a place — and the Music Agent turns it into style directions, lyrics, and a finished track. Custom mode and the editing tools are there when you want more control.

That's the point. Remix it into another style, extend it from the point you choose, or box-select a section on the waveform and rewrite just those bars — the lyrics stay time-aligned throughout.

Both, on the same canvas. Split any track into 2, 4, or 6 stems — vocals, drums, bass, instruments — and master it against one or more reference sounds, exported as WAV or MP3. To match a reference track of your own, use AI Mastering; for stem files you can take into a DAW, use the Stem Splitter.

Yes — tracks you generate are cleared for commercial use in videos, ads, podcasts, and monetized content. Pro plans include a downloadable commercial license certificate issued in your name.

It depends on the song you want to finish. Look for full songs with vocals, control over lyrics and style, and a way to fix the parts you don’t like. Lacuna lets you pick the model for each song — from the widest range of styles to lyrics sung exactly as written in the most vocal languages — and rewrite any section afterward instead of starting over.

Most AI song generators hand you an MP3 and stop. Lacuna treats the first song as step one: a canvas that keeps every version connected, with section editing, stem separation, mastering, analysis, and licensing on the path to release.

Your next song is one idea away

Generate the song, shape the arrangement, master it, and release it — free to start.