What is Suno Voices?
Voices is Suno's verified voice feature, introduced alongside the v5.5 model in March 2026. You give Suno a short recording of yourself singing, prove the voice is yours by reading a phrase aloud, and from then on your voice becomes a selectable option every time you generate a song. The model doesn't just imitate a tone - it performs your lyrics, in your voice, across genres you may never have sung in.
Songs that finally sound like you
Before Voices, every Suno track came back in somebody else's voice - a competent session singer who wasn't you. That's fine for a backing bed and useless for a demo. Voices closes the gap between "here's a song I generated" and "here's a song I made."
The practical unlock is range. You can hear yourself sing a genre you've never recorded, hold a note you can't hold, or deliver a lyric in a language you don't speak - all in a voice listeners will recognise as yours.
Your voice, any style
One recording covers everything. Generate bossa nova, drill, gospel or synthpop and the vocal stays recognisably yours across all of them.
Verified, not cloned
A spoken verification phrase confirms the voice belongs to you before it activates - so nobody can upload a recording of someone else and sing as them.
Private to your account
Only you can generate with your voice. You can publish songs that feature it, but the voice itself never becomes a public preset for other people.
Messy audio is fine
Acapella works best, but files with backing music are accepted - Suno automatically isolates the vocal with stem extraction before building the profile.
Works with your lyrics
Pair a Voice with Custom mode and your own words, and you get a finished performance of a song you actually wrote - not a generated approximation.
Adjustable influence
An Audio Influence slider controls how strongly the model leans on your recording. Keep it high when you want the resemblance to be unmistakable.
How to create your Voice
Six steps, and the whole thing takes about five minutes - most of which is finding a quiet room. You'll need a paid plan and an account belonging to someone 18 or over.
Find Voices in the Create panel
On the Create page, open the Add Voice dialog. If you've used Suno for a while, note that the Voices button has replaced Personas in this menu - Style Personas still exist, they now live inside the Voices menu rather than beside it.
Record live, upload a file, or reuse a Suno song
Three ways in: record straight from your microphone in the browser, upload an audio file you already have, or pull the audio from a song already sitting in your Suno library. All three end up in the same place.
Sing for 30 seconds to two minutes
Suno accepts anything from 15 seconds to 4 minutes, and asks you to pick the best two-minute segment. In practice, aim for 30 seconds minimum and 45-60 seconds for a solid profile. Sing something you know well, at the volume you'd actually use in a song - not shouting, not whispering.
Read the verification phrase aloud
Suno shows you a randomly chosen phrase and asks you to speak it. This is the anti-abuse step: it matches your speaking voice against the singing sample to confirm the voice is genuinely yours, and it's the reason you can't upload a recording of a famous singer and perform as them. You'll also confirm you have the rights to use the voice.
Fill in the profile (optional)
Give the voice a name, an image, and a skill level if you want to. None of it is required, but naming matters once you have more than one - a "full voice" profile and a "soft head voice" profile behave quite differently, and you'll want to tell them apart at a glance.
Select your voice and generate
Your voice now appears in the Voice dropdown on the Create page. Pick it, make sure v5.5 is the selected model, write your prompt or paste your lyrics, and generate. Push the Audio Influence slider high if the resemblance isn't strong enough.
How to record a sample that actually works
The quality of your Voice is decided almost entirely at the recording stage. A muddy sample makes a muddy voice, and no slider fixes it afterwards. Here's what separates a profile that sounds like you from one that sounds like a bad phone call.
Do this
- Sing acapella - your voice alone, no backing track. This is the single biggest quality factor.
- Give it 45-60 seconds. The floor is 15, but more material means a more stable profile.
- Sing something you know cold so you're not thinking about the words while you perform.
- Keep your volume consistent - the level you'd actually use on a record.
- Find a quiet, soft room. Curtains, a wardrobe, a car - anything that kills reflections.
- Cover your range - include a low phrase and a high one so the model hears both ends.
Avoid this
- Clips under ~20 seconds. They typically fail, and when they don't, the result is unstable.
- Singing over a track. Suno can strip the music out, but you lose fidelity doing it.
- Heavy processing - autotune, reverb, doubling. The model learns the effect, not you.
- Background noise - AC hum, traffic, a laptop fan, typing. All of it bleeds into the profile.
- Shouting or whispering. Extremes at either end skew how the model reads your tone.
- Anyone else's voice. Verification will catch it, and it breaks Suno's terms.
Why verification exists
Voice cloning is the part of AI music that most deserves suspicion, and Suno's design answers the obvious objection directly: you can only clone yourself. Here's exactly what that means in practice.
Liveness check, not a checkbox
The phrase is chosen at random and spoken in the moment, so a pre-recorded clip of someone else won't pass. It's a live proof of possession, not a box you tick.
Nobody else can use your voice
Your Voice is bound to your account. Other Suno users cannot select it, browse it, or generate with it - there's no public library of user voices.
You control what gets shared
Songs featuring your voice can be published like any other track. Whether others may remix or cover them is a choice you make at publish time.
Rights confirmation
Setup asks you to confirm you have the rights to the voice you're submitting. Uploading someone else's vocal is a terms violation regardless of whether it slips through.
Reportable in-app
If you believe a Voice misuses someone's likeness, it can be reported directly inside Suno rather than through a separate support process.
Watermarked like everything else
Tracks made with your voice carry the same inaudible AI watermark as any other Suno output - so provenance is still detectable downstream.
Voices vs Personas vs Custom models
Suno has three personalization features and they're easy to confuse, partly because Voices absorbed the Personas button in the Create menu. The cleanest way to hold them apart: Voices is whose voice, Custom models is whose style, Personas is which sound to carry forward.
| Feature | Answers | What you feed it | Best for | Plan |
|---|---|---|---|---|
| Voicesv5.5 | "Whose voice is singing?" | 15s-4min of you singing, plus a spoken verification phrase | Demos, covers of your own songs, hearing yourself in a genre you can't record | Pro & Premier |
| Custom modelsv5.5 | "Whose style is this?" | Six or more of your own finished tracks, used to fine-tune the model | Producers with a catalogue who want output that matches their sonic signature | Pro & Premier, up to 3 |
| Style Personas | "Keep this sound going" | An existing Suno song whose vibe you want to reuse | Building a consistent EP or series from one track you already like | Paid plans |
| My Taste | "What do I usually like?" | Nothing - it learns passively as you create | Better defaults and recommendations with zero setup | Free for everyone |
Where a Voice actually earns its keep
Voice cloning sounds like a novelty until you have a specific job for it. These are the ones that hold up.
Hear a finished arrangement of the track you just wrote, sung by you, without booking a session or setting up a mic.
Find out what you'd sound like as a country singer, a drill artist, or a 1970s soul vocalist before committing a day to trying it.
Send a producer or band a reference where the topline is already in your voice, instead of asking them to imagine it.
Birthday tracks, wedding songs, in-jokes - dramatically funnier and more affecting when it's genuinely your voice singing them.
Intros, outros and theme songs in your own voice give a channel a sonic identity that stock music never will.
Generate a guide vocal in your voice, then split the stems in Studio 2.0 and build the real arrangement around it.
Deliver a lyric in a language you don't speak, in a voice that still reads as yours to anyone who knows you.
People facing vocal surgery or age-related change use it to capture how they sound now, while they still sound like it.
Which plans include Voices
Voices is a paid feature. There's no trial of it on the free tier, so the entry price for putting your own voice into a song is the Pro plan.
- No Voices - not available on the free plan
- No Custom models
- My Taste personalization included
- 50 credits / day on v4.5-all
- Voices included - full verified voice creation
- Up to 3 Custom models trained on your catalogue
- v5.5 and the full model lineup
- 2,500 credits / month, commercial rights
- Everything in Pro, including Voices
- Suno Studio 2.0 - MIDI, effects, stem separation
- 10,000 credits / month
- Highest priority and concurrency
Troubleshooting
Four problems account for almost every failed Voice. All of them are fixable in a couple of minutes.
Suno Voices - frequently asked questions
The questions that come up most often once people have made their first Voice.
How long does my recording need to be?
Suno accepts 15 seconds to 4 minutes and asks you to pick the best two-minute segment. In practice, anything under about 20 seconds usually fails, and 45-60 seconds of clean acapella singing is the sweet spot. More material gives the model more of your range to work with, which is why a minute beats twenty seconds even when both are accepted.
Can I clone someone else's voice?
No, and that's deliberate. Setup requires you to read a randomly chosen phrase aloud, which is matched against your sample to confirm the voice is yours - a pre-recorded clip of somebody else won't get through. You also confirm you hold the rights to the voice you submit. Separately, prompts naming real artists are blocked. Voices exists to let you clone yourself, not anyone else.
Can other people use my voice?
No. Only you can generate with your Voice - it's bound to your account and doesn't appear in any public library. You can publish songs that feature it, and if you enable the option at publish time, other creators can remix or cover those songs. But that gives them the song, not the voice itself.
Is Voices free?
No - it requires a Pro or Premier subscription. Free accounts get My Taste personalization but not Voices or Custom models. Pro at roughly $8-10 a month is the cheapest route in, and it carries the same Voices functionality as Premier; Premier's extra is Studio 2.0 and a much larger credit allowance.
Does it work with every model?
No - Voices is a v5.5 feature. If you have an older model selected in the picker, your voice is ignored and you'll get a default singer. This is the single most common reason people think their Voice "stopped working," so check the model first whenever output sounds generic.
What happened to Personas?
The Voices button replaced Personas in the Create menu, but Style Personas weren't removed - they now live inside the Voices menu. They do different jobs: a Persona carries forward the overall sound of an existing song, while a Voice supplies the actual vocal identity. You can use both together.
Can I use a recording with music in the background?
Yes - Suno accepts files with backing music and automatically isolates the vocal using stem extraction before building the profile. But acapella still produces a noticeably better result. If the only recording you have is a mixed track, it'll work; if you can spend two minutes singing into your phone in a quiet room instead, do that.
Can I make more than one Voice?
Yes - the Add Voice dialog lets you create a new voice or pick from ones you've already made, and each can carry its own name, image and skill level. Multiple profiles are genuinely useful: a full-chest recording and a soft head-voice recording produce different singers, and naming them clearly saves confusion later.
Can I sell songs made with my voice?
Yes, provided you generated them while a paid plan was active - commercial rights on Pro and Premier cover tracks featuring your Voice just like any other output, and Suno takes no royalty. Two caveats carry over from all Suno music: songs made on the free tier stay non-commercial forever even if you upgrade, and a commercial licence isn't the same as a registered copyright.
Why can't I see the Voices option?
Three likely reasons. You're on the free plan (Voices needs Pro or Premier), the account belongs to someone under 18 (there's an age restriction), or your region isn't covered yet - geographic availability is limited but expanding. If none of those fit, contact support@suno.com.
Does my voice sound better in some genres?
Generally yes. Styles close to how you actually sang in the sample transfer most convincingly, while extreme distances - a soft-spoken sample pushed into full-throated metal, say - drift further from you. If a genre isn't landing, record a second Voice performing closer to that style rather than fighting the slider.
Can I combine Voices with Studio 2.0?
Yes, and it's the strongest workflow Suno currently offers. Generate a track using your Voice, open it in Studio 2.0 on Premier, split the vocal out with Advanced Split, then build a real arrangement around it - play parts in over MIDI, add effects, draw automation, and export 32-bit stems for your own DAW.
Your voice, every song you can imagine
Thirty seconds of singing and one spoken phrase is all it takes. Everything you make afterwards sounds like you.