Original score,
summoned from
a sentence.
An AI music generator for tabletop games. A new score for every scene.
Start summoningHow it works
From sentence to score
in four steps.
No prompt engineering. Generation takes about a minute.
Describe your scene
Write the scene in plain words.
A barrow at dusk. A banquet hall thick with smoke. Plain prose, the way you'd already write it in your prep notes. No tags, no parameters, no jargon.
Choose your mode
Pick the mode of play that frames it.
Did players roll for initiative or is someone seducing the handsome city mayor? Combat or roleplay dictates intensity.
Provide the context
Add the details.
Fantasy or sci-fi? What's the mood, the setting, the musical genre? Not sure about genre? SummonSong will make recommendations.
Summon
Take the track to the table.
Yours to keep. Yours to play in session. Yours to share back into the communal library so that another DM can call it up too.
Ten questions.
Three minutes. Then your pressing: which kind of DM you are, and exactly where to start.
Side B · The Library
Three things the library does
Every track another DM has summoned is in yours, too.
Search by scene, by mood, by who summoned it. The archive grows with every track.
Combat and Roleplay queued before initiative is rolled.
Tracks sort into mood playlists at the moment they're written. The right cue is one tap away, even when the party is already past the door.
Nothing scraped, nothing borrowed, nothing inherited.
Every track is composed from scratch for the prompt that made it. The streaming services, the lawyers, and the takedown bots stay outside the session.
Listen