How to use Suno: prompt to finished track
The hard part of Suno is not finding the buttons. It is four decisions: how detailed the prompt should be, whether you supply lyrics, how long to wait, and which take to keep.
Write the prompt in four slots
- Genre and era - city pop, 90s, lo-fi, orchestral.
- Mood - bright, restrained, tense, anthemic.
- Lead instrument - piano, 808, acoustic guitar, strings.
- Vocals - male, female, duet, or instrumental only.
One sentence covering those four beats a page of adjectives: "late-night Mandarin pop, piano intro, warm female vocal, no EDM" is already a usable prompt.
Lyrics: yours or the model's
You can let the model write them from a theme, or supply your own. If you write them, three rules matter: one line per musical phrase, section tags such as [Verse 1] and [Chorus], and no lines long enough to force the delivery to rush.
Waiting, and choosing a take
Submitting puts the task in a queue; audio usually arrives within a couple of minutes. Each generation returns two takes - different performances, not different lengths of the same thing. Keep the one with the stronger chorus and edit from there.
Export
MP3 for playback, uploads and sharing; WAV for editing and delivery masters, remembering that WAV is prepared on request and is much larger. Downloads are metered by plan, so do not use them as your audition button.
常见问题 / FAQ
How many songs does one generation return?
Two, usually finished within seconds of each other.
Can I edit one section instead of regenerating?
Yes - replace-section edits just that range and keeps the rest, which is cheaper and more consistent.
Does the interface language change the output?
No. Write the lyrics in the language you want sung; the UI language is irrelevant.
注册只需要邮箱，一次生成 0.6 元出 2 首，MP3 与 WAV 下载不限次。