Type an artist's name into Suno's Styles field and one of two things happens. Either Suno rejects the prompt, or it quietly ignores the name and gives you its idea of the genre. The first is a rule: Suno doesn't allow prompting with real artists' names. The second is the more useful lesson: a name was never an instruction. "Like Fleetwood Mac" tells a model which shelf the record sits on, not what's playing.
What you're actually asking for when you say "make it sound like…" is a set of musical decisions: a production era, a kind of voice, a drum feel, a bass part, two or three signature sounds and a way of mixing them. Suno can follow every one of those. So the job is to translate the name into the decisions.
Why naming an artist doesn't work, even when it slips through
Two reasons, beyond the rule.
An artist isn't one sound. Most careers cover several eras, producers and moods. "Sounds like Prince" could mean a dry drum machine and falsetto, or a full band and a guitar solo. The model has to guess, and it guesses the average.
Names pull in clichés, not specifics. The things people say about a famous artist ("iconic", "legendary", "funky") are exactly the evaluative words that make Suno's output generic. You want the opposite: the parts a session band would need to hear before they play.
There's also a practical one. Prompts that lean on someone else's identity, especially their voice, are the ones most likely to run into Suno's filters and streaming services' impersonation rules when you release. Describing the sound in your own words keeps the song yours.
Decode the sound in six parts
Listen to the record you have in mind and write one short phrase or sentence for each of these. Order matters: it's the order Suno weights Styles in.
- Genre lead tag and tempo. The closest genre in Suno's own wording, then BPM. "nu-disco, 118 BPM". If you don't know the tempo, a BPM-tap app on your phone gets it in ten seconds.
- Era or production colour. One phrase, not a list: "late-70s Californian studio rock, warm tape". This does more than any adjective.
- The lead vocal. Timbre, register, delivery, and whether it's stacked: "airy male falsetto, close and breathy, doubled on the hook".
- Drums. What they play and how they feel: "loose swung kick and snare, crisp shuffling hats, slightly behind the beat".
- Bass. The part and its job: "rubbery synth bass with slides, syncopated, locked to the kick".
- Two to four signature parts and the mix. The sounds you'd miss if they were gone ("detuned Rhodes chords, chopped soul samples as stabs"), then where things sit ("vocal close and dry, low end up front, pads wide").
Then list what Suno tends to add that this sound doesn't have, and put those in Exclude, never as "no …" in Styles: "trap hi-hats, EDM risers, distorted guitars".
Three worked examples
Here's what that looks like for three references people ask about a lot. Each one started as the output of Session Sheet's Sound like… feature (more on that below), trimmed for length. None of them contains a name.
"Sounds like Kaytranada"
future funk, 112 BPM. 2010s Montreal beat-scene house, swung and soulful. Smooth R&B vocal, laid-back and breathy, with chopped vocal hooks. Loose swung kick and snare with off-grid clap, crisp shuffling hats, slightly behind the beat. Rubbery sub-heavy synth bass, bouncing syncopated lines that lock to the kick. Detuned Rhodes chords with warm chorus, chopped soul vocal samples as rhythmic stabs. Warm, punchy low end up front, vocal close and dry.
Exclude: EDM risers, trap hi-hats, rigid quantized drums, distorted guitars
"Sounds like Fleetwood Mac, the Rumours era"
soft rock, 108 BPM. Mid-70s Californian studio rock, warm analog polish. Dry, tight kit in a steady mid-tempo pocket, crisp snare backbeat, soft tom rolls into choruses. Melodic fingered electric bass walking under the chord changes. Fingerpicked acoustic guitar arpeggios driving the verses. Vocals forward and dry with a short room, harmonies panned wide.
Exclude: heavy distortion, synth leads, EDM risers, autotune
"Sounds like Burial"
dubstep, 138 BPM. Mid-2000s South London 2-step, nocturnal and rain-soaked. Pitch-shifted, chopped female vocal fragments, wistful and ghostly. Shuffled 2-step kick and snare with skipping hats, loose and slightly muffled. Deep, round sub bass with sparse, slow-moving notes. Vinyl crackle and tape hiss bed, rain and city noise, soft detuned pad chords. Vocal fragments buried in reverb, drums dusty and low.
Exclude: wobble bass, EDM risers, big drops, four-on-the-floor kick
Notice what each one does. The era phrase carries half the identity. The vocal sentence describes a voice type, never a person. And every Exclude list names what v6 would otherwise reach for: risers and drops on the dubstep, distortion on the soft rock.
Put it to work in a prompt
The description isn't the whole Styles field. It's the sound. A full prompt still needs this song's arrangement (sections, with bar counts for the intro) and an ending, in the order from our v6 prompt guide. Keep the total around 800 characters; the free Styles checker shows the count, and flags any evaluative or negative words that crept in.
If you want the sound in a different genre ("that vocal and those keys, but over UK garage drums"), keep the vocal, parts and mix sentences, and swap the genre, tempo, drums and bass for the new genre's. Our genre prompt pages give you the drums and bass for 90 genres to drop in.
Two things not to do
Don't upload someone else's track as audio. Suno's audio features are for your own recordings, and other people's songs are what its copyright filter is there to catch.
Don't aim for a specific singer's voice. Describe a voice type, not a person: a soft, breathy alto is a sound; "her voice" is an impersonation. Streaming services' rules on AI releases are strictest about exactly this. More in our release checklist.
Let Session Sheet do the decoding
In Session Sheet, the song form has a Sound like… box. Type an artist, a song or an era, press Describe it, and you get the era, vocal, drums, bass, parts and mix as editable sentences, plus a suggested genre and tempo and the Exclude terms. Any name is stripped out before you see it, and it's never saved or sent to Suno. The description then shapes the Styles the app writes for your song, alongside your own genre, lane and lyrics. It costs one studio action, and your first songs are free.