Glossary

The words the writer and this guide use, in the order you meet them.

Patch
One complete result: the three Suno boxes (Lyrics, Style of Music, Exclude styles), the slider settings, and an analysis of the piece.
Engine
What writes the patch. The song engine sets lyrics as a sung, dark, electronic pop song; the track engine builds an instrumental-first electronic track from a brief. How the writer picks one.
Brief
A few words about a mood, a place or a scene, as opposed to lyrics. A brief makes a track.
Knob
Any control in the writer that changes the patch.
Family
One of 12 song genres, from minimal art-pop to futurepop. Picking one makes a song.
Style
One of 103 electronic genres in 25 scenes, from Berlin techno to liquid drum & bass. Picking one makes a track.
Flavour
A sub-type of a style, for the seven styles that cover several scenes. Flavours.
Era
The decade a track sounds like, from the 1980s to the 2020s. Eras.
Arc
The energy shape of the piece over time: where it builds, peaks and drops.
Seed
The number that selects one variation. The same text, knobs and seed always give the same patch; another seed gives a sibling.
Layer
One part of a patch you can reroll on its own. Arrangement: sections and the energy path. Sound: instruments, groove, textures and mix. Wording: the words in the tags and the spoken lines. Each has its own seed; 0 follows the main seed. Style, BPM, key and voice stay with the main seed.
Mutation
0–100: how far the patch may drift from its family or style. At 0 it draws only from its own palette; higher values borrow from neighbouring ones.
Tag density
1–5, from bare to maximal: how much detail the section tags in the Lyrics box carry.
Hook and topline
The hook is the phrase a track sings; the topline is the shape in which it is sung, from a single chant to every stanza. Hooks.
Variation bank
Eight sibling patches from the same text and knobs, side by side, to compare and pick from.
Library
The patches you saved, kept in this browser. Rate them after trying them in Suno.
Taste
A lean on what Auto picks, learned from your ratings: families and styles you rated well come up more often.