To get exactly what you want out of Suno, you need to turn on Custom Mode. This splits the prompt into two distinct boxes: "Style of Music" and "Lyrics."
In Suno, getting the right genre and vocal style comes down to mastering two things: the Style Prompt (for the overall vibe) and Metatags (for structure and vocal delivery).
1. The "Style of Music" Prompt (The Vibe)
The Style box dictates the genre, tempo, instruments, and mood of the entire track.
The Golden Rule: Use 15 to 30 comma-separated keywords.
Instead, build a layered prompt using this formula:
[Primary Genre], [Subgenre], [Vocal Style], [Key Instruments], [Mood/Vibe], [Production Style]
Examples:
Acoustic/Folk:
Indie Folk, Americana, husky male vocal, acoustic guitar, upright bass, intimate, warm, tape-saturatedEDM:
Deep House, female falsetto vocal, heavy sub-bass, analog synthesizers, driving, euphoric, 120 BPM, club mixCinematic:
Cinematic Orchestral, operatic female choir, lush string section, huge timpani, epic, building, dramatic
What to Avoid:
Do not use contradictory terms (e.g., "fast slow ballad" or "happy melancholic")—this confuses the AI.
2. Using Metatags (Structure and Vocals)
Metatags are commands written in [square brackets] inside the Lyrics box.
Structuring the Song
Always place structure tags on their own line immediately before the lyrics they apply to.
[Intro]/[Outro]: Starts and ends the track.[Verse 1],[Verse 2]: Where the story unfolds.Keep verses to 4-8 lines. [Pre-Chorus]: Builds tension leading into the main hook.
[Chorus]: The main, catchy hook. Keep this short (2-4 lines) for maximum impact.
[Bridge]: Adds a new melody or chord progression in the second half of the song.
[Guitar Solo]/[Instrumental Break]: Forces the AI to stop singing and focus on the instruments.
Controlling Vocal Style
You can use metatags to change how the AI sings a specific section. Add these tags right below your structure tags.
| Tag | Effect | Best For |
| [Whisper] | Soft, intimate delivery | ASMR, intimate ballads |
| [Spoken Word] | Talking rhythmically | Intros, storytelling, Hip-hop |
| [Belting] | Powerful, sustained high notes | Pop choruses, musical theater |
| [Harmonies] | Layered backing vocals | Pop, Gospel, R&B |
| [Growl] | Aggressive, distorted vocal | Metal, hard rock |
Example of a fully formatted lyrics box:
[Verse 1]
[Soft, intimate female vocal]The neon lights are buzzing low
Walking down a street I used to know
[Pre-Chorus]
[Building energy]The shadows stretch across the wall
[Chorus]
[Belting, layered harmonies]And I'm caught in the middle of the night!
Waiting for the morning light!
3. The Iterative Workflow
Don't expect perfection on the first try. AI music is about iteration.
Generate 2-3 variations: Listen to how the AI interprets your prompt.
Adjust the tags: If the AI sings a metatag out loud, simplify the wording (e.g., change
[Really loud epic guitar solo]to just[Guitar Solo]).Extend: If you get a great first minute but the song cuts off, use Suno's "Extend" feature. You can select the exact timestamp where the song sounds best and generate the next section from there.
