Your AI song came out wrong: what to change
AI song not right? Edit your prompt. Add instruments, tempo, mood, or genre. Specify what you want and what to exclude for better results.
If your AI-generated song came out wrong, change your text prompt. Adding more specific details about instruments, tempo, mood, or genre will help Suno produce a track closer to your vision. You can also tell the model what to exclude.
Make your prompt specific
The more specific your prompt, the better the result. A prompt like a pop song is too broad. The model has to guess at many elements. Instead, describe the exact elements you want to hear.
Add instruments
Specify the instruments you want. Think about the lead instrument, rhythm section, and any melodic or harmonic elements. For example, instead of just electronic music, try:
An upbeat electronic dance song with a strong synth bassline, arpeggiated lead synths, a four-on-the-floor drum beat, and a bright pad sound.
Define the genre and subgenre
Genres like "rock" or "pop" cover a huge range of sounds. Try to be more precise. If you want a rock song, what kind of rock? Is it classic rock, punk rock, alternative rock, or something else? Similarly, for pop, is it synth-pop, K-pop, or acoustic pop?
A lo-fi hip-hop beat with a smooth jazz sample, a relaxed drum loop, and a deep bassline.
Describe the mood and tempo
Mood words like "happy," "sad," "energetic," or "calm" help the model understand the emotional tone. Tempo can be described with words like "fast," "slow," "upbeat," or using beats per minute (BPM) if you know it.
A melancholic classical piece with a slow tempo, featuring a solo cello, a mournful violin melody, and a gentle piano accompaniment.
Specify vocal styles and lyrics
If your prompt includes vocals, describe the style. Do you want male or female vocals? What kind of voice: a high-pitched soprano, a deep baritone, a whispered spoken word? You can also provide specific lyrics within your prompt, enclosed in quotation marks.
An indie folk song with a female vocalist singing in a soft, airy voice. Lyrics: "Sunrise paints the window pane, another day begins to wane."
Use negative prompts to exclude elements
Sometimes it's easier to say what you don't want. Suno supports negative prompting, where you can specify elements to exclude. Use the NO keyword followed by the element.
An ambient electronic track with swirling pads and a gentle rhythm. NO drums, NO vocals.
This tells the model to create an ambient track but specifically avoid any percussive elements or singing. You can also specify instruments or genres to avoid.
A soulful R&B track with a strong groove. NO acoustic guitar, NO trap drums.
Understand model variability
Suno uses advanced AI models that generate music probabilistically. This means that the same prompt can yield different results each time you generate. If a song isn't quite right, try generating it again with the exact same prompt before making changes. You might get a better variation.
Copyright and commercial use of AI-generated music
Copyright in AI-generated output is currently unsettled. Various copyright offices, including the U.S. Copyright Office, have stated that works without human authorship may not be registrable. You should check the current copyright laws in your country regarding AI-generated content. Suno tracks carry an inaudible SynthID watermark, indicating they are AI-generated.
For commercial use, a paid Suno plan is required. The free plan does not include commercial usage rights. Please refer to our terms of service for full details on usage rights and licenses.