SKILL.md — Writing Effective Suno v5 Prompts

Overview

Suno v5 has two input fields: a Style Prompt and a Lyrics/Structure Field. The style prompt defines the sound. The structure field defines the shape. Both matter, but they do different jobs. This document covers how to write both for any genre, any mood, any use case.

The Two Fields

Style Prompt

This is a short block of text (keep it under ~200 characters for reliability) that tells Suno what the track should sound like. It controls genre, tempo, instruments, mood, production quality, and exclusions.

Lyrics / Structure Field

This is where you put either actual lyrics with section tags, or (for instrumentals) section tags with performance cues and punctuation patterns to shape the arrangement. Structure tags go in square brackets on their own line.

Style Prompt Anatomy

A good style prompt has 4-6 components in this order:

[Genre + Subgenre], [BPM], [Key (optional)]. [Instruments]. [Mood/Emotion]. [Production/Mix cues]. [Exclusions].

1. Genre + Subgenre

Be specific. Suno was trained on real music and responds to real genre vocabulary. The more precise you are, the tighter the output.

Bad: "electronic music" Good: "melodic techno" Best: "deep melodic techno, progressive structure, Berlin club sound"

Bad: "rock" Good: "post-punk revival" Best: "post-punk revival, angular guitars, driving motorik beat"

You can blend genres with commas or "meets" / "x" / "with":

• "trip-hop meets orchestral, downtempo cinematic"
• "Afrobeat x jazz fusion, complex polyrhythms"

2. BPM

Always include a specific number. Suno respects BPM fairly well.

• Ambient/downtempo: 60-90
• Hip-hop/R&B: 80-100
• Pop/indie: 100-120
• House/techno: 120-130
• Drum and bass/jungle: 130-175
• Drill'n'bass/breakcore: 160-200

3. Key (optional but useful)

Specify if you care about harmonic content or plan to layer/stitch multiple generations. Use standard notation: "C minor", "Eb major", "A minor", "F# minor".

Minor keys sound darker, more tense, more introspective. Major keys sound brighter, more resolved, more euphoric.

4. Instruments

Name 3-5 specific instruments. This is one of the most powerful levers. Don't say "synths" — say which synths. Don't say "guitar" — say what kind.

Generic: "synths, drums, bass" Specific: "analog Juno pads, 303 acid bassline, chopped amen breaks, sub bass"

Generic: "guitar and piano" Specific: "fingerpicked nylon guitar, upright piano, brushed snare, upright bass"

Suno responds to instrument-specific vocabulary:

• "Rhodes piano" ≠ "grand piano" ≠ "honky-tonk piano"
• "reese bass" ≠ "sub bass" ≠ "303 acid bass" ≠ "fuzz bass"
• "gated reverb snare" ≠ "tight dry snare" ≠ "brushed snare"

5. Mood / Emotion

Use 2-3 adjectives that describe the feeling, not the sound. Suno is surprisingly good at translating emotional language into musical choices.

• "melancholic but hopeful"
• "tense, building, cathartic release"
• "playful, irreverent, slightly chaotic"
• "vast, lonely, beautiful"

Avoid vague superlatives: "epic", "amazing", "cool" mean nothing to the model.

6. Production / Mix Cues

These tell Suno how the track should be mixed and mastered. Think of them as instructions to an engineer.

Useful cues:

• "clean mix, clear instrument separation"
• "lo-fi, warm tape saturation, vinyl crackle"
• "wide stereo, heavy reverb, spacious"
• "dry, punchy, in-your-face, compressed"
• "large hall reverb on snares"
• "tempo-synced delay"
• "analog warmth, slight tape wobble"

7. Exclusions (Negative Prompts)

Tell Suno what NOT to include. Use "no [element]" format. Critical for instrumentals.

• "no vocals, no singing, no humming, no choir"
• "no autotune"
• "no guitar solo"
• "no EDM drops"
• "no trap hi-hats"

Keep exclusions to 2-3 max. Too many can hollow out the arrangement.

Style Prompt Examples

Ambient / Drone

Dark ambient drone, 72 BPM. Granular textures, evolving pad layers, field recordings,
deep sub bass. Meditative, vast, unsettling calm. Reverb-heavy, wide stereo, no rhythmic
pulse. Instrumental only, no vocals.

Jazz

Modern jazz trio, 140 BPM, Bb major. Upright bass walking lines, ride cymbal swing,
acoustic piano comping with bebop voicings. Sophisticated, conversational, lively.
Clean live recording feel, room ambience, no compression. Instrumental only, no vocals.

Hip-Hop

Boom-bap hip-hop, 92 BPM, D minor. Dusty vinyl-sampled drums, Rhodes piano chops,
deep sub bass, scratched vocal stabs. Nostalgic, confident, head-nodding groove.
Lo-fi mix, subtle tape saturation, punchy kick. No autotune, no trap hi-hats.

Orchestral / Cinematic

Cinematic orchestral, 66 BPM, C minor. Legato string ensemble, French horn swells, harp
arpeggios, timpani rolls, solo cello melody. Sorrowful, sweeping, building to grandeur.
Wide dynamic range, concert hall reverb, no percussion until second half. Instrumental only.

Punk / Garage

Garage punk, 178 BPM, E major. Distorted power chords, driving floor tom beat, fuzz bass,
shouted unison vocals. Raw, urgent, sloppy-on-purpose energy. Lo-fi recording, room mic
bleed, no polish, no reverb. Short, under 2 minutes.

Synthwave

Synthwave, 100 BPM, A minor. Analog arpeggiated bass, gated snare, chorus-drenched
guitar stabs, Juno pad washes, retro lead synth. Nostalgic night-drive mood, neon melancholy.
Tape warmth, sidechain compression, wide stereo pads. No vocals.

Structure Field: Section Tags

Place tags in square brackets on their own line. These define the arrangement.

Standard Song Tags

[Intro]
[Verse] or [Verse 1], [Verse 2]
[Pre-Chorus]
[Chorus]
[Bridge]
[Instrumental Break]
[Outro]

Extended / Specialized Tags

[Build]
[Drop]
[Breakdown]
[Ambient Break]
[Percussion Break]
[Drum Break]
[Climax]
[Fade Out]
[Hook]
[Solo]

Performance / Delivery Tags (placed before or within sections)

[Whispered]
[Belted]
[Spoken Word]
[Building]
[Powerful]
[Gentle]
[Half-time]
[Double-time]

Structure Field: Instrumental Tracks

For instrumentals, you still use section tags but replace lyrics with either:

1. Parenthetical descriptions — production notes in plain English inside parentheses
2. Punctuation patterns — lines of dots, exclamation marks, and spaces that nudge Suno toward rhythmic/percussive content rather than vocals

Parenthetical descriptions are more reliable. Punctuation patterns are less predictable but can help force instrumental passages and influence energy.

Example: Instrumental with descriptions

[Intro]
(sparse piano, single notes, large reverb, building tension)

[Verse]
(drums enter, steady groove, bass joins, melody develops)

[Chorus]
(full arrangement, all instruments, maximum energy, anthemic)

[Bridge]
(stripped back, just bass and drums, tension rebuilds)

[Chorus]
(final chorus, even bigger, added harmonies and layers)

[Outro]
(instruments drop out one by one, ending on sustained piano chord)

Example: Instrumental with punctuation

[Intro]
. . .
. .

[Build]
.. .! . .. ! .
. ! .. .

[Drop]
!! .! !! .! !! .! !! .!
.!! !! .!! !! .!! !! .!!

[Breakdown]
. ! . . ! .
. . .

[Outro]
.
.

Combining both approaches

[Intro]
(single instrument, ambient, building)
. . .

[Drop]
(full energy, all instruments, driving rhythm)
!! .! !! .! !! .!

[Outro]
(fading, sparse, resolved)
.

Dynamic & Narrative Arcs

The biggest difference between a forgettable AI track and a compelling one is dynamic arc — the track needs to go somewhere. Suno v5 can follow arc instructions if you embed them in both the style prompt and the structure field.

In the Style Prompt

Describe the journey in one sentence:

• "Dynamic arc: quiet ambient intro building to full-intensity peak, then graceful breakdown"
• "Emotional arc: melancholic verse, hopeful pre-chorus, euphoric chorus, reflective outro"
• "Energy arc: starts sparse and tense, builds steadily, peaks at 2/3, resolves quietly"

In the Structure Field

The section tags themselves define the arc. Order matters. A common mistake is making every section the same energy level. Instead, think in terms of energy percentage:

[Intro]       — 10% energy
[Verse 1]     — 30% energy
[Pre-Chorus]  — 50% energy
[Chorus]      — 80% energy
[Verse 2]     — 40% energy
[Bridge]      — 20% energy (contrast!)
[Final Chorus] — 100% energy
[Outro]       — 15% energy

Use parenthetical cues to reinforce this:

[Bridge]
(strip everything back, just vocals and a single instrument, quiet, vulnerable)

Multi-Section / Stitched Approach

For longer or more complex pieces, generate each section separately and stitch them in Suno Studio or a DAW. This gives you:

• Tighter control over each section's energy and texture
• Ability to iterate on one section without regenerating everything
• Consistent key/tempo across sections (set explicitly in each prompt)
• Better transitions (crossfade or hard-cut to taste)

When stitching, keep these consistent across all prompts:

• BPM (exact same number)
• Key
• Core instrument list
• Production style descriptors

Change these between sections:

• Energy descriptors
• Which instruments are active
• Mood adjectives
• Mix intensity

Vocal Prompting (When You Want Vocals)

In the Style Prompt

Specify register, tone, and delivery:

• "Male baritone, warm, slightly raspy, restrained emotion"
• "Female alto, breathy, intimate, close-mic feel"
• "Choir, ethereal, layered harmonies, cathedral reverb"

In the Structure Field

Use delivery tags inline:

[Verse 1]
[Whispered] In the silence of the night
[Building] I feel you pulling me close
[Belted] AND I WON'T LET GO

[Chorus]
[Powerful] We rise together
Through any weather

Tips:

• ALL CAPS text tends to be sung louder/more intensely
• Short lines = punchier delivery
• Longer lines = more flowing/legato phrasing
• Hyphens in words extend them: "lo-ove", "sooo-long"
• Specify rhyme scheme if it matters: add a note like "(AABB rhyme scheme)"

Iteration Strategy

Suno v5 is stochastic — the same prompt produces different results each time. The best workflow is:

1. Start rough. Short style prompt, basic structure. Generate 3-4 takes. Listen for which ones capture the right feel, even if details are wrong.

2. Identify what's working. Note which elements sound right: the drum feel, the bass tone, the overall energy arc, the vocal delivery.

3. Refine one variable at a time. Change only one thing per iteration so you can hear what each change does. "dark" → "ominous" → "foreboding" will each give different shadings.

4. Use negative prompts to fix problems. If unwanted elements appear (random guitar, background vocals), add targeted exclusions.

5. Generate 4-6 takes of your final prompt. Pick the best. Small variations between takes are normal and desirable — you're curating, not dictating.

6. Use Suno Studio for surgery. If a take is 90% perfect but has a bad section, use region editing to regenerate just that section rather than the whole track.

Common Pitfalls

Too many words. Suno responds better to crisp, specific phrases than to long prose descriptions. "Warm analog synth pads, wide stereo" beats "I want the synth pads to sound warm and analog, spread across the stereo field with a nice wide image."

Contradictory instructions. "Aggressive, calm" or "minimal, dense" confuse the model. Pick a direction. If you want contrast, put it in the arc ("calm verses, aggressive choruses"), not in a single descriptor.

Overloading exclusions. More than 3-4 negative prompts can make the output feel hollow or confused. Exclude only the things that keep showing up unwanted.

Ignoring structure. A style prompt alone gives you a formless blob of sound. The section tags are what give it shape, dynamics, and narrative. Always use them.

Expecting exact reproduction. Suno is generative, not deterministic. It interprets your prompt, it doesn't execute it literally. Write prompts that describe a space of acceptable outputs, not a single exact track. Then curate from multiple takes.

Using artist names. Suno blocks direct artist/band references. Instead, describe the sonic qualities you associate with that artist. "Radiohead" → "experimental art-rock, complex time signatures, atmospheric guitars, falsetto vocals, melancholic beauty."