How to Use Your Voice In Suno: How to clone your voice in suno

IS YOUR VOICE CLONE IN SUNO : https://www.youtube.com/watch?v=zyshKhOCIyA

Or view dozens of HOW TO USE SUNO Tutorials https://www.youtube.com/watch?v=zyshKhOCIyA

Suno Voice Cloning Actually Works? How to Create a Persona From Your Own Voice

Stop using external apps! In this Suno AI tutorial, I show you the native workaround to clone your voice in Sunodirectly. Unlike other methods that require Controlla or Kits AI, this guide teaches you how to upload your vocals, extract stems, and create a custom Suno Persona entirely within the platform

If you've ever wanted Suno to sing with something that feels like your voice (without needing perfect pitch or a studio setup), there is a surprisingly workable path. It is not a perfect carbon copy, and it does not replace a real vocal recording. Still, it can produce a highly stylized version of your voice that carries over some of your phrasing and musical habits, which can be the difference between a generic AI singer and something that feels personal.

This post walks through a practical workflow: record a short vocal, upload it as a sample, generate a stylized vocal, then save it as a Persona using a native workaround inside Suno. After that, you can create full songs using the Persona and refine results by tightening your prompts.

If you are brand new to Suno and need an account first, the creator provides a Suno sign-up link in the video description.

What "voice cloning" in Suno really means right now

The term "voice cloning" sets expectations high, so it helps to define what you are likely to get. In this workflow, Suno does not recreate an exact copy of a singer's voice. Instead, it tends to generate a version that is "you, but processed," like a character performance based on your input. That can still be valuable if your goal is to keep a consistent identity across songs.

The most useful outcome shown here is that Suno can pick up on how you sang the sample, not just the fact that it was a voice. That includes things like phrasing, the way notes connect, and the overall vibe of your delivery. In other words, the result can sound like a stylized persona that carries your fingerprints, even if the tone is not identical to your raw vocal.

One more reality check: the feature set and behavior change often. Settings that work one week may need slight prompt tweaks the next. That is why the process below focuses on a repeatable workflow you can adjust, rather than a single magic prompt that never changes.

If you want broader context on how Suno frames voice cloning as a concept, Suno also maintains its own write-up at Suno's AI voice cloning guide.

Record and upload a short vocal sample

The starting point is simple: make a short recording of your voice and upload it as the audio sample Suno will reference.

In the walkthrough, the sample is recorded directly using Suno's audio recording option, then uploaded. The clip is roughly 15 seconds and includes singing (not just speaking). That matters because Suno is being asked to produce a singing vocal, so giving it sung material helps it latch onto your performance style.

After recording:

1. Upload the clip and confirm that Suno should use it as the sample.
2. Clear the lyrics box (remove any words that might steer the generation away from the sound test).
3. In the Style field, describe the kind of vocal you want Suno to generate.

That last part is easy to overlook. Even with an uploaded sample, Suno still reacts strongly to the written style description. If your natural voice sits lower or higher than typical, being explicit about voice type can prevent Suno from drifting into a different range.

For example, the walkthrough calls out a common issue: a low voice getting interpreted as a high male vocal. To counter that, the style is specified as a female alto in an indie rock direction.

If you are brand new to Suno and need an account first, the creator provides a Suno sign-up link in the video description.

The settings used to get a stylized match (and why iterations matter)

Once the sample is in, the next step is dialing settings so Suno follows the uploaded audio instead of wandering off into a generic singer. The walkthrough uses Advanced Options and keeps control set to Manual, then intentionally reduces some of the more "wild" behavior.

The two sliders called out directly are:

• Weirdness set all the way down
• Style Influence initially set all the way down (then later raised)

There is also a numeric control that gets pushed into the red, landing around 87. The key detail is not the exact label, but the behavior: pushing that value higher changes the output, and the creator prefers it in the high 80s for this use case.

To keep track of attempts, the generation is labeled (for example, "test one"). That sounds minor, but it becomes helpful fast because this process often takes several runs. In the demo, it takes about six iterations to get to a result that feels right.

Here is a compact snapshot of the configuration shown:

Setting / Input	What was used in the walkthrough	Why it matters
Vocal sample	About 15 seconds of sung audio	Gives Suno a performance reference
Lyrics box	Cleared for the initial test	Avoids lyric content driving the output
Style text	"female alto vocal indie rock"	Steers range and genre
Advanced options	Manual	Keeps behavior more predictable
Weirdness	All the way down	Reduces unexpected changes
Style influence	Down at first, later around 30	Balances the sample vs the written style
Numeric slider	High 80s (around 87)	Changes how strongly the model behaves in this mode
Attempts	Multiple generations	Helps you converge on a usable voice

The big takeaway is that the first generation is not the finish line. In this workflow, iteration is part of the process, especially when Suno updates.

The prompt change that finally worked

After several attempts, the prompt gets simplified and the results improve.

The key changes:

• The style prompt becomes: "female alto vocal indie rock"
• The word "sample" is removed from the style prompt
• A simple lyric, "la", is added

That "la" choice is practical: it gives Suno something to sing without adding lyrical content that might distract from evaluating the voice. On the second iteration after the prompt change, the output follows the sample more closely and sounds like a stylized version of the original singer.

How to save your voice as a Suno Persona (native workaround)

After you get a generated vocal that feels close enough, the next goal is turning it into a Persona you can reuse. The twist is that using the normal Persona creation path may not work directly from an uploaded audio clip, so the workflow uses stems as a bridge.

The workaround looks like this:

1. On the successful generation, open the menu (the three dots).
2. Choose Get stems, then extract the vocal and instrumental.
3. Wait for stems to generate (no download is required for this workflow).
4. Return to the Create area, open filters, and uncheck "Hide stems".
5. Locate the vocal stem (the "la" vocal in this case).
6. Open the three dots on that vocal stem, choose Create, then Make persona.
7. Name the Persona so you can recognize it later.
8. Set it to private by unchecking public, then save.

Once saved, that Persona becomes a reusable vocal identity you can select when generating new songs. The important detail is that the Persona is being created from the extracted vocal stem, not from the original upload..

Create a full song using your Persona (then refine the range)

With a Persona saved, you can move from voice testing into actual songwriting. In the walkthrough, the next step is entering full lyrics, selecting an indie rock direction again, and generating a complete song with the Persona.

The process stays mostly the same, with one big difference: now you are judging musical performance in context. A voice that sounds good on "la" can still feel off once it has to carry real lines, real melodies, and a full arrangement.

The first test song output is close, but the vocal comes out too high. Instead of changing everything, the adjustment is targeted: the prompt is updated to tell Suno to sing in a low range. That single clarification makes a noticeable difference, and the next song aligns better with the intended sound.

This is the part that tends to feel most like working with a human vocalist. You are not only choosing a genre, you are also directing performance choices. The more direct the request (female alto, indie rock, low range), the easier it is to steer away from the default pop-soprano tendency many generators drift toward.

If you want a related project that also involves recording vocals into Suno, check out this tutorial : recording vocals into Suno project tutorial.

What this approach is best for (and what to expect next)

At the end of the workflow, you have something useful: a Persona that sounds like a stylized version of you, and a repeatable process for generating songs that keep some of your identity. The most interesting result is not perfect realism, it is consistency. Suno picks up on the way you sing, which can make AI-made songs feel less like stock demos and more like tracks that belong to one artist.

It also sets expectations clearly. This is about getting "you, interpreted," not "you, duplicated." If your goal is a flawless vocal replica, you may find the current limits frustrating. On the other hand, if you want a signature vocal feel for original songs, this workflow is already productive.

Because Suno changes quickly, treat this as a baseline process: upload, simplify prompts, iterate, then lock in the best take as a Persona through stems. From there, prompt tuning (like controlling range) becomes the main tool for getting closer to what you hear in your head.

For readers who want to follow more of the creator's work beyond this tutorial, the channel also shares updates and extras through a monthly email list for AI music news and prompts.

Conclusion

If you're trying to bring your own identity into Suno songs, this Persona workflow is a solid way to get there. The results will sound like a stylized version of your voice, but they can still carry the phrasing and feel that make your singing recognizable. The stem-based workaround is the key that turns a good generation into a reusable Persona. From there, small prompt edits, like asking for a lower range, can be the difference between "close" and "that feels right."