Updated
Suno Structure Tags That Actually Work: [Verse], [Chorus], and Friends
Square-bracket tags in the lyrics box tell Suno how your song is built. How they work, which ones matter, and the tag-stacking trick that improves how often Suno follows them.
Open the lyrics box in Custom mode and you’ll see people writing things like [Verse 1] and [Chorus] above their lines. Those square-bracket structure tags are how you tell Suno the architecture of your song, where sections begin, what kind of section each one is, and how the energy should move.
Used well, they’re the difference between a song with a real shape and a four-minute ramble. They work like this.
The core vocabulary
These are the workhorse tags, in the order a typical pop song uses them:
[Intro]
[Verse 1]
[Pre-Chorus]
[Chorus]
[Verse 2]
[Pre-Chorus]
[Chorus]
[Bridge]
[Chorus]
[Outro]
Each tag goes on its own line, in square brackets, right before the lyrics of that section. Suno reads them as structural guidance: verses get verse-like energy, choruses lift, bridges shift, outros wind down.
Two habits that help: number your verses ([Verse 1], [Verse 2]) so repeated sections stay distinct, and keep labels consistent, if you call it [Chorus] once, don’t switch to [Hook] later for the same section.
Tags are strong suggestions, not commands
Important expectation-setting: structure tags guide the model; they don’t bind it. Suno usually respects a clear, conventional structure, but it can merge sections, stretch them, or occasionally ignore a tag, especially in unusual arrangements. If a generation ignores your structure, regenerate before rewriting; often the next roll follows it fine.
What makes tags more likely to be followed: a conventional overall shape, lyrics whose length matches the section (two lines of chorus lyric won’t sustain a 40-second chorus), and a Styles field that isn’t fighting the structure. That middle point deserves emphasis — lyric quantity is itself a structure signal. Rough starting points: verses 4–8 lines, pre-choruses 2–4, choruses 4–8, bridges 2–6. And paste repeated choruses in full at every occurrence rather than trusting the model to infer the repeats: the worked example below does exactly this.
The stacking trick
Here’s a technique worth stealing: you can put multiple tags on the same line, and it works better than spreading them across separate lines. This is how you attach vocal instructions to a specific section:
[Verse 1] [Female Vocal] [Soft]
Quiet opening lines here...
[Chorus] [Layered Harmonies] [Powerful]
Big chorus lines here...
Stacked on one line, the tags read as a single instruction for that section, and in repeated use and community reports, section-level compliance improves versus listing each tag on its own line. It’s a strong pattern rather than guaranteed syntax: the model can still ignore, reinterpret, or occasionally even sing an unusual tag, so test anything important across a couple of generations.
Beyond the basics
A few more tags that earn their keep:
[Instrumental Break], a section with no vocals; describe it if you want, e.g.[Instrumental Break: guitar solo][Intro: instrumental]— makes clear the opening has no singing[Outro]followed by[End]— nudges the song toward an actual ending instead of an awkward fade[Ad-libs]— background vocal flourishes, best used sparingly
Resist the urge to invent elaborate custom tags for everything. The model responds best to conventional section names; exotic tags mostly get ignored.
One punctuation rule worth burning into memory: square brackets are for instructions, parentheses are for singing. Anything you put in parentheses gets treated as lyrics to perform, which is exactly what you want for echoed backing lines like (ooh, tonight) and exactly what you do not want for stage directions. Write (quietly now) and there is a good chance a voice will sing the words “quietly now.” If it is a direction, put it in square brackets; if it should be heard, put it in parentheses. This is the single most common tag mistake in the wild.
A complete worked example
[Intro: instrumental]
[Verse 1] [Soft]
First verse lyrics...
[Pre-Chorus] [Building]
Rising lines here...
[Chorus] [Powerful] [Layered Harmonies]
The big hook...
[Verse 2]
Second verse lyrics...
[Pre-Chorus] [Building]
Rising lines again...
[Chorus] [Powerful] [Layered Harmonies]
The big hook again...
[Bridge] [Stripped Back]
The turn...
[Chorus] [Powerful]
Final hook...
[Outro]
[End]
Pair a structure like this with a clear Styles field — see the Styles field guide, and you’ve covered the two inputs that matter most in all of Suno.