SoMusicPowered by AiX Society
繁简EN
Module 3 · Hands-on

Suno AI Workshop

Suno shrinks the distance between “one sentence” and “a complete song” to almost nothing. It won’t replace you, but it turns an afternoon of arranging into thirty seconds — if you know how to prompt it, and which step must come back to you.

Back to home/Suno AI Workshop
ModelAccessPosition & best useStatus
v6
FlagshipPredictable, precise
The most stable and most faithful to your prompt. Use it once you know the direction and need a usable result. Section rewriting, line-by-line lyric replacement and multi-source mixing all live here.
v6-wild
ExperimentalUnpredictable, varied
Deliberately keeps the “accidents” and imperfections — great for collecting unexpected directions at the ideation stage, then bringing your pick back to the flagship model.
v6-mini
LightweightFast, free
The only model available on a free account. Fast and light, but the result sounds more obviously AI-generated — good for validating an idea, not as a final product.
Free

The v6 family launched on 9 Sep 2026, retrained on licensed Warner Music, BMG and Believe catalogs; v4 through v5.5 have been retiring. Work you made with older models won’t disappear — it can still be played and downloaded.

Prompt generator

Suno’s “style” field is most stable in English. Pick from the five groups below and instantly assemble a style prompt you can paste straight in.

1 · Genre
2 · Mood
3 · Vocals
4 · Arrangement anchor (one or two instruments is enough)
5 · Lyric language

Paste this text into Suno’s “Style” field. Lyrics go in the “Lyrics” field, and section tags go inside the lyrics field, one per line.

Section tags: take the arranging back

Suno can structure a song itself, but it usually writes the “most common pop structure”. To make it follow you, write the sections in square brackets in the lyrics field — the lowest-cost, highest-return move.

[Intro] [Verse 1] [Pre-Chorus] [Chorus] [Verse 2] [Chorus] [Bridge] [Final Chorus] [Outro]
[Intro]Intro: give texture first[Verse 1]Verse: narrative, detail[Pre-Chorus]Pre-chorus: build heat[Chorus]Chorus: a one-line hook[Bridge]Bridge: change perspective[Final Chorus]Final chorus: the peak[Outro]Outro: finish by subtraction[Instrumental]Instrumental: let the singer breathe

Advanced: add a note after the tag, e.g. [Chorus - full band, stacked harmonies], to control a section’s density.

The five most common mistakes

Common approachWhy it failsWhat to do insteadImpact
Writing “a nice song”
Prompt too vague (the model can only guess) 
Make at least three of genre, mood, instruments and vocals specific. The more the description feels like a sensory experience, the closer the result to the sound in your head.
Tags in the wrong field
Sections crammed into the style field (different rules) 
The style field only reads timbre and genre, not structure. Section tags go in the lyrics field, one per line.
Settling after one generation
No selection (each run is a sample) 
Generate the same prompt four to six times, keep the best one, then tweak a line or two with section rewriting instead of redoing the whole song.
Releasing the raw output
Skipping post-production (loudness off-standard) 
Export stems into a DAW and remix, or at least re-master. Streaming platforms generally reference −14 LUFS.
Using the free tier for commercial work
Wrong license (can’t be fixed) 
The free plan is personal, non-commercial only. For commercial use you must generate during a paid period — upgrading later does not back-apply licensing to existing songs.

Plans, downloads & commercial use

PlanMonthlyWhat you getCommercial use
Free
Free (no card needed) 
v6-mini model, 50 credits/day (~10 songs); a lifetime total of seven downloads; no extra credits available.
Pro
$10 / mo (~$8 yearly) 
v6 and v6-wild, 2,500 credits/mo (~500 songs), up to 30 minutes of upload audio, Voices and up to three custom models, basic stems; 20 downloads/mo.
Commercial use OK
Premier
$30 / mo (~$24 yearly) 
10,000 credits/mo (~2,000 songs), full Suno Studio 2.0 (MIDI, automation, plugins, 32-bit stem export), advanced stems; 60 downloads/mo, unlimited via Studio export.
Commercial use OK

Three license traps that are easiest to step into

  • Songs made on the free tier cannot be used commercially. YouTube monetization, client projects and ad music all count as commercial use — even as background music.
  • Upgrading later does not back-apply licensing. Songs generated during a free period don’t become commercial just because you paid later; regenerate commercial work during an active subscription.
  • Credits don’t roll over and downloads are capped. Subscription quota doesn’t carry across months, and downloads have plan limits — plan your flow before large output or you’ll get stuck at the worst moment.

Compiled from public information as of September 2026. Suno’s prices, quotas and terms change often — verify against suno.com before adopting. Also: v6 was trained on licensed catalogs, but the Sony Music and Universal Music lawsuits are still ongoing; assess commercial risk yourself.

Cantonese & Suno: the best use today

This is the most practical part of the course. Suno’s Cantonese diction is still improving, but tone-to-melody alignment isn’t yet at a level you’d safely release.

What worksArrangement & melody drafts

Get a complete arrangement in thirty minutes to communicate “this is the direction I want” to an arranger, singer or client — extremely efficient.

Watch outThe Cantonese vocal

Cantonese’s six tones must be compatible with the melody, and AI doesn’t always get it right yet — it can sound like it’s “singing the wrong word”.

Suggested flowSuno for the draft, humans for the final

Use Suno for the arrangement and melody direction, then swap the vocal track for a human performance. You keep the speed and the tonal accuracy.

Do it nowWrite lyrics and tags fully

Fill the lyrics field with Cantonese and one section tag per line — diction improves noticeably, and this step is completely free.

What Suno can’t do (as of now): no third-party VST/AU plugins, no MIDI output to your DAW (Premier’s Studio can export MIDI); complex jazz harmony and polyrhythms are still weak; cloning named artists’ voices is banned for legal reasons.