Prompt
A short description of the sound: genre, mood, instruments and tempo, written as plain words such as 80 BPM in A minor. Prompt mode also writes the words for a sung track, so every word you type should be one you would miss if it were gone.
Brief · two takes · keep one
AI Music Gen
An AI music gen run starts from a short written brief and returns two full tracks you can play side by side. Name the genre, the mood, two instruments and a tempo, decide whether you want vocals or an instrumental bed, then listen to both takes on the same headphones and keep the one that already sounds closest to the song in your head.
- 01BriefGenre, mood, two instruments, tempo
- 02Two takesEach run gives two tracks
- 03KeepOne clear winner, one note on why
What an AI Music Gen Run Lets You Decide
You write the ask; the tool writes the performance. Four choices shape every run.
Own Lyrics
Switch to Own Lyrics and paste lines you already wrote. Your lines stay fixed while the melody and groove are built around them, which suits a name, a slogan or a verse you love.
Instrumental Only
Switch it on when a voice-over, a game scene or a study session needs a purely instrumental bed. The brief stays the same; only the sung part drops out.
Inspiration tags
Tags such as Country, Folk, Rock, Blues and Disco give a quick start when the box is empty. Pick one, then add your own mood and instruments so the result carries your idea.
Number of Songs starts at 2, so each run gives two tracks, and songs can run up to 8 minutes. That pair is the real feature: two readings of one brief tell you more about your prompt than any single file could.
Four Steps Through an AI Music Gen Session
A loop you can finish in one sitting: write, generate, compare, then stop or change one thing.
- 1
Write a tight brief
One genre, one mood, two instruments and a tempo. A clear four-part brief beats a paragraph of stacked adjectives.
- 2
Choose voice or bed
Paste your lines in Own Lyrics or let Prompt mode write the words for a sung track, or switch on Instrumental Only when speech will sit on top.
- 3
Hear both takes
Play the pair back to back at the same volume. Write one sentence on what each take does better than the other.
- 4
Keep or change one thing
Download the clearer take, or change a single word in the brief and run it again so you know what moved.
Stop after a second run unless you can name the exact gap that is left. A named gap, such as a chorus that should lift higher, gives the next prompt a job.
AI Music Gen Prompts and the Takes They Made
Most AI gen music sounds generic for one reason: the brief was generic. Four slots fix most briefs: genre, mood, two instruments and a tempo.
Both tracks below came from briefs built that way, one sung with pasted lines and one with Instrumental Only switched on. Play them as drafts to learn from.
River Lights · sung, own lyrics“Bright city pop, 112 BPM, slap bass, electric piano chords, crisp drums, warm female vocal”, with a short verse and chorus pasted in Own Lyrics mode.
Reading Room Bed · Instrumental Only“Warm lo-fi hip hop, 80 BPM, muted Rhodes chords, brushed drums, soft upright bass, calm and steady, room for a spoken voice.”
Judge music generated by AI like a demo: in the sung take, check that the chorus words land on the strong beats; in the bed, check that the midrange leaves room for a voice. A words to music generator also works from a prompt alone, such as “acoustic folk, hopeful, fiddle, leaving a small town”. Swap one detail at a time.
Where an AI Music Gen Track Ends Up
Most runs serve one of four jobs. Naming the job before you write the brief keeps the prompt short.
A bed under a cut
Name the tempo so an editor can cut on the beat, and switch on Instrumental Only when a narrator will talk over it.
A demo of your lyric
Paste the verse and chorus you wrote and hear two melodies for them before you book a session or pick up a guitar.
A loop for a long session
Ask for a calm, steady bed with few changes, so it sits quietly behind study, gaming or a long work block.
A song for one person
Write a prompt with a name, a place and a shared memory, and keep the take that sounds most like them.
If a job needs both a sung version and a bed, write one brief and run it twice, once with vocals and once instrumental, so the two files share a sound and you can use both in the same project.
Who Reaches for an AI Music Gen
People open a music gen AI tool for three different reasons, and each one rewards a short brief.
A
Video and podcast makers
A bed that matches the script's tone and leaves room for speech.
- Name the tempo so cuts land on the beat
- Switch on Instrumental Only and keep one file per episode
B
Songwriters with finished lines
Two melody sketches for words that are already written.
- Switch to Own Lyrics and paste the verse and chorus
- Keep the phrasing you like and rewrite the rest by hand
C
Curious first-timers
To hear what an artificial intelligence music generator does with one line.
- Start from a one-line mood or an Inspiration tag
- Add instruments and tempo one run at a time
Whichever group you are in, the habit is the same: write a note on what changed between runs and keep only the take you would play again.
AI Music Gen Habits That Save Reruns
Small edits to the brief make side-by-side listening fair and keep each new run useful.
One mood, two sounds
Pick one mood and two concrete instruments. Five stacked adjectives hide which word moved the result.
Change one thing
Alter the tempo or one instrument between runs, one edit at a time, so you can hear what each edit does.
Play the second track
Listen to both takes all the way through before you rewrite. The keeper is often the one you nearly skipped.
Keep every brief in one running note with the date and the take you kept. With gen AI music, that note soon becomes your best guide to what works for your ear.
Where an AI Music Gen Does Its Best Work
A fair scope lets a good draft be judged as a draft. The tool handles the first take; you handle the final call.
Strong atTurn a short brief into two full tracks per run, with vocals or as an instrumental bed.
Your partJudge the mix on headphones and speakers before a track goes out.
Strong atSing your own lines in Own Lyrics mode, or write the words for you in Prompt mode.
Your partBring original ideas, your own lyrics and recordings you hold the rights to.
Strong atReturn songs that can run up to 8 minutes.
Your partWrite a specific brief: concrete words return a focused take on every run.
Treat the AI music gen output as a strong first draft. Keep the take that works, then trim, extend or remix it in the editor you already use.
AI Music Gen Questions Before You Generate
Short answers to what visitors ask before the first run.
Do I need to write my own lyrics?
1/5
Lyrics are optional. You can paste lines you wrote in Own Lyrics mode, describe the song in Prompt mode and let the run write the words, or switch on Instrumental Only and skip singing altogether. Pasting your own lines makes sense when the exact words matter, such as a name or a slogan. Prompt mode works when you care more about the feeling. Start with a prompt if you are unsure, then paste your best lines into the next run.
Why does each run give two tracks?
2/5
Two readings of one brief show you what your words controlled and what the run invented. If both takes share a tempo and instruments, the prompt did its job; if they differ wildly, the brief left too much open. Listen to both at the same volume, write one sentence on each, and use that sentence to decide the next edit. Many keepers turn out to be the take you played second.
How long can a generated song be?
3/5
Songs can run up to 8 minutes. If you need a long bed for a stream or a study session, ask for a calm, steady arrangement with few changes, so the track stays in the background. For a song with a story, name the sections you want, such as verse, chorus and bridge, so the song has a clear shape. Any tempo or key you want goes into the prompt as words, for example 72 BPM in D major.
What do I get on the free plan?
4/5
The free plan gives you 30 credits a year, about 15 generations, at 2 credits per song; unused credits reset on your renewal date. Free users can download a limited number of songs, while paid plans download in MP3, WAV and FLAC. For more runs on a one-time basis, you can also buy credits once with Pay Once. Spend early runs on short, clear briefs, so each one teaches you something about your prompt.
Can I put a generated track in a client video?
5/5
Monthly and yearly plans both include commercial rights, within the scope of the Commercial License. The Commercial License certificate costs 5 credits per song on monthly plans and Pay Once, and 0 credits on yearly plans, with no extra charge. Keep the certificate with the audio in the project folder, so you can show where the music came from if a client or a platform asks.
Still unsure? Run one short brief with vocals and the same brief as an instrumental, then compare all the takes side by side.
Next step
Start Your Next AI Music Gen Run
Write one genre, one mood, two instruments and a tempo, then keep the clearer of the two takes you get back.
- BriefFour slots, one change per rerun
- KeepOne take, one note on why