Suno AI prompts: a formula and examples by genre

A strong prompt is the difference between a track that sounds like everything else and one that sounds like what you meant. Suno does the singing and the playing, but it only follows the brief you give it, so the words you type are the real instrument. This guide breaks down a Suno prompt formula that works, then hands you Suno prompt examples by genre you can copy, adjust, and build on.

What a Suno prompt actually is

A Suno prompt is a short music brief, not a wish list. The model reads it as a set of instructions and fills every gap you leave, so anything you do not decide, it decides for you. That is why two people typing "happy song" get wildly different results, while a precise brief lands in the same place run after run. The goal is not more words. It is the right words, in an order the model can follow.

How Suno reads your prompt

It helps to picture what happens to the words after you press create. Suno does not match your text to a library of songs. It generates new audio that fits the description, weighting the terms it recognises most strongly, which is usually the genre and the vocal. That is why a clear genre anchors the whole result, and why a mood word on its own tends to drift. Terms the model cannot place get diluted, so a prompt stuffed with clever adjectives often sounds less focused than a plain one with four solid choices. Write for the parts the model leans on hardest, and the rest falls into line.

A prompt formula that works

Most reliable prompts follow the same order: genre, mood, key instruments, tempo, the vocal, and a note on production. Four to six of these is the sweet spot. Fewer, and the output drifts toward generic. More, and the model starts ignoring parts of the request.

Assembled, that reads: warm lo-fi hip hop, relaxed and nostalgic, dusty piano and soft boom bap drums, 75 BPM, no vocals, vinyl crackle. Every part answers a question the model would otherwise guess, which is what keeps the result steady.

Suno prompt examples by genre

These are copy ready starting points. Paste one into the style box, then change a single element to hear what it controls. Each keeps the formula above so you can see the pattern.

• Pop: bright modern pop, hopeful, jangly guitars and claps, 120 BPM, airy female vocal, polished and radio ready.
• Lo-fi: warm lo-fi hip hop, relaxed and nostalgic, dusty piano and soft drums, 75 BPM, no vocals, vinyl crackle.
• Hip hop: hard trap, confident, booming 808s and crisp hi hats, 140 BPM, gruff male rap, dark and spacious.
• Rock: driving alt rock, defiant, distorted guitars and live drums, 145 BPM, raw male vocal, big and energetic.
• EDM: festival house, euphoric, sidechained synths and four on the floor kick, 128 BPM, chopped female vocal, wide and bright.
• Cinematic: cinematic orchestral, tense and building, strings, brass and timpani, slow tempo, instrumental, epic and wide.
• Country: modern country, warm and easygoing, acoustic guitar, fiddle and brushed drums, 95 BPM, heartfelt male vocal.
• R&B: smooth 90s R&B, sultry, electric piano and soft bass, 85 BPM, silky female vocal, intimate and lush.
• Metal: melodic metal, intense, palm muted guitars and double kick drums, 160 BPM, powerful male vocal, heavy and tight.
• Folk: acoustic folk, gentle and reflective, fingerpicked guitar and light strings, 90 BPM, soft male vocal, close and honest.
• Ambient: ambient chill, calm and floating, warm pads and field recordings, very slow, instrumental, spacious and soft.

More genre starting points

The same formula stretches across styles the first list did not cover. Here are a few more to paste and adjust.

• Synthwave: retro synthwave, moody and nostalgic, analog synths and gated drums, 100 BPM, instrumental, neon and wide.
• Drum and bass: liquid drum and bass, uplifting, rolling breakbeats and warm bass, 174 BPM, soulful female vocal.
• Reggae: classic reggae, laid back and sunny, offbeat guitar, organ and deep bass, 80 BPM, warm male vocal.
• Blues: slow electric blues, aching, expressive guitar and Hammond organ, 65 BPM, gritty male vocal.
• Gospel: uplifting gospel, joyful, piano, organ and full choir, 100 BPM, powerful lead and harmonies.
• Afrobeats: modern afrobeats, warm and danceable, syncopated percussion and bright synths, 108 BPM, smooth vocal.

Prompting for instrumentals

Not every track needs a voice, and instrumentals are often easier to get right. Add no vocals or instrumental to the prompt, and lean harder on the instruments and the mood, since the melody now carries the whole song. For background music behind a video or a podcast, describe the feel you want under the talking, such as calm, unobtrusive, warm pads and soft piano, and keep the arrangement simple so it does not fight the voice on top. Instrumentals also tend to survive looping better, which matters when the track sits under content that runs longer than the song.

Prompting for a mood or a scene

Sometimes you are chasing a feeling rather than a genre, and you can prompt that directly. Picture where the music will play and describe it. A study playlist wants calm, steady, and distraction free. A workout track wants driving, high energy, and a strong beat. A sad scene in a video wants sparse, slow, and a single lead instrument. Naming the use often produces a better fit than naming a genre, because it tells the model what the music has to do rather than only what it should sound like. You can still add a genre and a tempo on top, and the two together give a tight, purposeful result.

Structure your lyrics with tags

In custom mode you can paste your own lyrics, and Suno reads simple tags that tell it where each part of the song goes. Put each tag on its own line.

• [verse], [chorus], [bridge] and [outro] mark the sections, and a marked chorus usually returns a stronger hook.
• Round brackets add production notes inside a line, such as (echoing harmonies), (guitar solo), or (whispered).
• Short, singable lines beat long, dense paragraphs, since the model has to fit your words to a melody.

A quick example of the shape: a [verse] tag, four short lines, a [chorus] tag, two lines you want to stick, then a [verse] again. Keeping the strongest line in the chorus and keeping it short gives the hook the best chance of landing.

Advanced prompt techniques

Once the basics work, a few sharper moves give you more control over the result.

• Name the vocal precisely. Soft breathy female or gritty baritone steers the take far more than a word like good.
• Describe dynamics. Phrases like slow build into a big final chorus or drops to just piano at the bridge shape the arrangement.
• Exclude what you do not want. Saying no vocals or no drums is often as useful as asking for something.
• Set a tempo and, if it matters, a key. A number gives the model a target where fast or slow leaves it guessing.
• Keep style and lyrics separate. Put the sound description in the style box and the words in the lyrics field, rather than mixing them.

Naming styles the right way

How you reference a style changes how well it lands. Naming a decade or a scene works, since 90s R&B, UK garage, or 80s synth pop point the model at a clear era. Naming a single famous artist works far less well, because the result is inconsistent, and publishing a track that imitates a real singer is a legal risk you do not want. If you have a specific artist in mind, describe what you actually like about them instead, such as the breathy vocal, the sparse production, or the walking bass, and you get closer without borrowing a name.

Working within the style box

The style field is not endless, and that is a feature. A short, ordered prompt uses the space better than a long one, so lead with the terms that matter most in case the tail gets trimmed. Commas are enough to separate ideas, and you do not need full sentences, since the model reads a list of descriptors as cleanly as prose. If you find yourself writing three lines of style, that is usually a sign to cut back to the four or five choices that define the track.

Change one thing at a time

The fastest way to improve a prompt is also the most patient. When a take is close but not right, change a single element and generate again, rather than rewriting the whole thing. Say the vocal is good but the energy is low. Raise the tempo from 90 to 110 and leave everything else alone, and you learn exactly what that number did. Do the same with one instrument, one mood word, or one production note. After a few rounds you stop guessing, because you have watched each term move the result on its own. It is slower than a rewrite for one attempt and much faster across ten.

Common prompt mistakes

Most weak results come from the same handful of habits.

• Overloading the prompt with ten adjectives, which makes the model drop half of them.
• Leaving the vocal undefined, so it changes on every run.
• Writing the mood but no instruments, which returns a vague bed of sound.
• Chasing a named artist's voice, which is unreliable and a poor idea for anything you plan to share.
• Rewriting the whole prompt after a bad take instead of changing one element and learning from it.

A copy and paste starter pack

If you just want something that works right now, take one of these full prompts and run it, then tweak.

• Uplifting indie pop, warm and hopeful, jangly guitars, punchy drums and bright synths, 118 BPM, airy female vocal, polished.
• Late night lo-fi, mellow and nostalgic, dusty Rhodes piano, soft boom bap drums and vinyl crackle, 72 BPM, no vocals.
• Epic cinematic trailer, tense and rising, staccato strings, deep brass and pounding timpani, slow build, instrumental.
• Smooth soul ballad, tender and intimate, electric piano, warm bass and light strings, 80 BPM, silky female vocal.

Save the ones that work as your own templates, and start the next song from a small edit of a template rather than a blank box.

From a prompt to a finished song

A prompt gets you a strong opening, and the rest is shaping. When a generation is close, use the extend tool to grow it into a full arrangement rather than starting over, and keep the same style phrase so the new sections match. If the melody is right but the mix is off, a remix pass with a revised prompt can fix it while keeping the shape. Treat the first result as a draft, listen to both takes, and save the version you like before you push further, since a promising take is easy to lose once you experiment on top of it.

A quick prompt cheat sheet

To keep it in one place, a workable prompt names the genre, the mood, one or two key instruments, a tempo in BPM, and the vocal, in that order, with a short production note at the end if you have one. Keep it to four to six choices, be specific about the voice, exclude what you do not want, and change a single element between runs. Save anything good the moment it lands. Follow that and most of your generations come back usable, which leaves your credits for refining a track rather than rescuing it.

Saving the songs your prompts make

A good prompt is only worth as much as the file you keep from it. Once a run lands, download the track so a strong take is not lost when a session fills up. The free Suno downloader turns a Suno link into an MP3 or WAV in a couple of clicks, with no account, which is the fastest way to save a keeper before you move on to the next idea.

If you are still finding your way around the tool, our guide on how to use Suno AI covers simple and custom mode, and the one on how to download Suno songs covers formats and devices. Prompt, listen to both takes, save the one you like, then refine.

Frequently asked questions