Suno prompts with a producer's style formula

Good Suno prompts put the sound in the Style field as one comma-separated line, with the most important descriptor first. Use this 13-slot order: genre, fusion genre, tempo, key, mood, mood, instrument, instrument, instrument, production, production, vocal, language. Around 12 descriptors is comfortable; past about 20, later words often start arguing with earlier ones.

The Style field should describe the record, not the plot. A line like make a sad song about leaving town with a huge chorus and rain in the lyrics asks for sound, story, structure, and words all at once. That usually creates a mushy result: the genre is vague, the chorus may not grow, and the lyrics can drift.

Use Style for sonic decisions. Put the actual scene, title phrase, and section map in Lyrics. If you start with a plain-language idea, shape the story first with text-to-song prompts, then reduce the Style field to sound words.

The fix is not longer. It is cleaner. Story words moved to Lyrics, and the sound stayed in Style.

The safest order is genre first, then fusion genre, tempo, key, two moods, three instruments, two production cues, vocal, and language. Early words tend to steer harder, so do not hide garage rock, trap soul, or boom bap after reverb settings.

<genre>, <fusion genre>, <tempo>, <key>, <mood>, <mood>, <instrument>, <instrument>, <instrument>, <production>, <production>, <vocal>, <language>

Copy the shape, not every word:

1 garage rock, 2 power pop, 3 156 BPM, 4 E major, 5 urgent, 6 bright, 7 distorted rhythm guitar, 8 driving bass, 9 splashy drums, 10 dry production, 11 small room reverb, 12 raw male vocal, 13 English vocals

The fusion genre is optional. When it helps, it should narrow the target: garage rock, power pop says rough guitars with a sharper chorus. For English lyrics, you can skip the language slot unless diction is a problem. For other languages, state the language near the end. If you need style starting points, browse genre ideas for AI songs and translate what you like into this order.

Strong Suno style prompts choose one clear value from each axis instead of stacking near-synonyms. Tempo is the obvious trap. 92 BPM is a direction. slow, fast, driving is a tug-of-war. A useful exception is 140 BPM, half-time feel, because the number sets the grid while the feel tells the vocal pocket to sit slower.

For an instrumental, write instrumental in Style and remove every lyric line.

Do not use artist names; describe the era, drums, hook instrument, mix shape, and vocal tone instead. Phrases such as in the style of and sounds like point at imitation, and they also give the model less useful build instructions than a plain sound recipe.

Ask yourself what you actually hear in the reference. Is the snare gated or dry? Is the hook a synth, a guitar riff, or a group chant? Is the vocal breathy, low, shouted, or choir-like?

This works because it tells the generator what to build, not who to copy.

Section tags belong in the Lyrics field, never inside the Style field. Keep each tag on its own line. If a tag is misspelled or placed beside a lyric, the model can sing it as a word. It sounds silly. The fix is boring, which is good.

Use a clean skeleton like this:

[Intro]

[Verse 1]
First image line
Second image line

[Pre-Chorus]
Short lift line
Short lift line

[Chorus]
Main title phrase
Main title phrase returns

[Bridge]
A new angle
A shorter answer

[Outro]
Main title phrase

[End]

• Tags are on their own lines
• No lyric line appears before the first tag
• The chorus text is identical each time it returns
• The outro has a clear last phrase
• The Style field has no bracketed tags
• Instrumentals have no lyric lines hiding after a tag

For a fuller tag reference, use the dedicated Suno meta tags guide. Keep this page focused on style prompting.

A complete prompt pairs one tight style line with lyrics that support the same groove. Input idea: late-night bus ride, regret, clean female vocal, synth pop, chorus should feel bigger than the verse.

Style
indie synth pop, new wave, 104 BPM, A minor, melancholic, wistful, analog synth arpeggios, soft drum machine, jangly electric guitar, 80s production, room reverb, warm female vocal, English vocals

Lyrics
[Intro]

[Verse 1]
Neon slides across the glass
Your last text is fading blue
I count the stops I should have passed
Still riding back to you

[Pre-Chorus]
The city hums under my shoes
I almost call, then I refuse

[Chorus]
I missed my stop again tonight
I saw your face in every light
The rain keeps writing on the window
I missed my stop, I let you go

[Verse 2]
Coins shake loose inside my coat
The driver turns the radio low
I hear the words I never wrote
And watch the empty crosswalk glow

[Chorus]
I missed my stop again tonight
I saw your face in every light
The rain keeps writing on the window
I missed my stop, I let you go

[Outro]
I missed my stop, I let you go

[End]

Expected result: a mid-tempo, minor-key synth pop track where the verse stays restrained and the chorus opens with wider guitars or pads. If it sounds too happy, change only wistful to lonely. Leave the BPM, key, and lyric alone for that pass. If you already have words, start from a lyrics-to-song setup and polish the Style field after the lyric structure feels solid.

When a prompt misses, change one descriptor, run two more generations, and compare before touching anything else. Same prompt, same lyrics, one changed axis. That is the only way to tell whether the vocal, production, or tempo actually helped.

The tempo math is simple: 104 BPM minus 8 percent is 95.68, rounded to 96 BPM. Plus 8 percent is 112.32, rounded to 112 BPM. If the two A versions differ more from each other than A differs from B, that axis probably is not the main issue. Move on.

On SoundUser, the same discipline helps when you describe a song, pick a style, and work with Lyria 3.5 in the AI song generator.

Most prompt problems come from late genre words, missing vocal cues, uneven lyric lines, stray vocals in instrumentals, vague tempo, or weak endings. Do not rewrite the whole song after one bad pass. Fix the likely cause and listen again.

For verse, chorus, bridge, and outro planning before you prompt, read song structure basics.

A prompt note turns each song into useful information by saving the style line, the one change you tried, and what changed in the output. Keep it short. You are building taste, not a giant spell book.

Style A:
Variant B one change:
What improved:
What got worse:
Keep for next song:

House defaults should be compact: boom bap, 90 BPM, C minor, gritty, tense or country rock, 98 BPM, D major, warm, nostalgic. Do not paste the same full tag-list into every track. Store one preferred tempo range per genre, such as 90-96 BPM for relaxed boom bap or 120-128 BPM for dance-pop, then adjust for lyric density.

Final pass: genre first, one fusion genre, exact BPM, optional key, two moods, three instruments, two production cues, one vocal cue, one language cue if needed.