Suno v5.5 prompting guide

Six techniques that separate "generic AI track" from professional-sounding output. No fluff.

1. The prompt formula
2. Style vs Lyrics — two control zones
3. Left-to-right: token priority
4. 5-8 tags — the sweet spot
5. One-variable revision method
6. Common mistakes

1. The prompt formula

Treat it as a short production brief, not a pile of adjectives. Each slot controls one sound layer. Drop a slot and Suno guesses the default — usually not in your favour.

Weak: sad piano song
Strong: cinematic pop ballad, melancholic and hopeful, 72 BPM, warm male vocals with slight reverb, piano and strings driven, building to a soaring chorus, for short-form creator content

2. Style vs Lyrics — two control zones

Suno v5.5 has two brains, and they listen to you differently:

Field	Role	Bracketed tags
Style	Global DNA — genre, mood, instrumentation, production	Baseline influence
Lyrics	Local director — switches states section by section	~10× stronger

Want a chorus with a guitar solo? Don't put it in Style — put it in Lyrics:

[Chorus | High Energy | Anthemic | Electric Guitar Solo]
Rise up — break free — let the whole block hear you roar...

3. Left-to-right: token priority

Suno reads the Style field left-to-right and weights the first 1–3 tokens most heavily. Whatever comes first becomes the "genre anchor" defining everything after.

dark trap, 808s, aggressive ≠ aggressive, 808s, dark trap

First gives mid-tempo trap with dense bass. Second gives an aggressive mix of nothing in particular — "aggressive" alone has no concrete musical semantics.

4. 5-8 tags — the sweet spot

A focused prompt with 5-8 strong, non-conflicting tokens consistently beats a chaotic one with 15+. Strong = specific instrument name, BPM number, or production texture. Weak = "beautiful", "epic", "cool".

Weak token	Strong replacement
powerful drums	thunderous taiko + roomy acoustic kit
warm synths	Yamaha DX7 electric piano + Roland Jupiter-8 brass stabs
cool vocals	raspy male vocal with throat-shredding belt on chorus
fast tempo	140 BPM driving

5. One-variable revision method (debug)

If a generation misses — don't rewrite the whole prompt. Change exactly one variable and compare:

• Wrong energy? — change tempo / energy descriptor
• Too cluttered? — remove one instrument
• Vocal off? — change gender / character of vocal anchor
• Doesn't fit platform? — add use case (short-form / workout / podcast)

6. Common mistakes

1. Overload. «Cinematic EDM pop rock with trap drums and orchestra» — модель потеряет якорь.
2. Vague words. «Beautiful», «epic», «vibey» — это для людей, не для AI.
3. No use case. Трек для тренировки и для подкаста — это два разных промпта, даже в одном жанре.
4. Artist names. Suno их фильтрует. Поэтому в нашем каталоге их нет — описываем сигнатуры через продакшн-приметы.
5. Style-only. Игнорировать Lyrics-поле = терять 90% контроля над аранжировкой.