Skip to main content
Professional Voice Clone (PVC) builds a studio-quality voice from recordings you provide. You upload your audio, verify that the voice is yours, wait for training, preview the result, and then release the voice. When you release it, you can earn a share of what other creators pay to use it.
Looking to make an original character voice instead? See Voice Design.

​
Before you begin

You must have permission to clone the voice: you own it, or you have the speaker’s consent. You confirm this before you release the voice.
  • Sign in at fish.audio/app.
  • A paid plan with a free PVC slot. Professional Voice Clone is not included in the Free plan. Each paid plan includes a number of PVC slots, and the Create Voice hub shows how many you have used.
  • Time. The full flow takes about 1 to 2 hours. Training alone takes 10 to 30 minutes.
  • Recordings. At least 10 minutes of clean audio of one speaker.

​
Open Professional Voice Clone

Go to fish.audio/app/create-voice and choose Professional Voice Clone. In the Create Professional Voice Clone? dialog, click Confirm. Fish Audio creates the clone and opens its workspace.
The Create Professional Voice Clone confirmation dialog
If you already have an unfinished draft, the dialog also offers Continue draft. If you have used all your slots, a dialog asks you to upgrade your plan.

​
Create the clone

1

Upload your audio

Add recordings of the voice you are cloning.
  • Formats: MP3, WAV, FLAC. Other common audio and video files are also accepted.
  • Each clip: less than 1 minute.
  • Recommended: 12 to 15 clips, each 45 to 60 seconds, with a consistent tone. Make sure the clips add up to at least 10 minutes.
  • Total: at least 10 minutes. The timeline shows your progress toward the minimum (up to a 180-minute maximum).
The Professional Voice Clone upload area with supported formats, clip guidance, and the 10-minute to 180-minute timeline
Each file shows its duration, size, and a status as it is checked. If a clip is rejected (for example, “No usable speech detected…”) or flagged for multiple speakers, delete it and upload a better one.
2

Analyze

When you have at least 10 minutes of usable audio, click Analyze Audio. Fish Audio checks the quality and language of each clip. When the analysis is done, click Continue.If you change your clips after the analysis, the results are cleared and you must analyze again.
The upload page with 14 uploaded clips, 12 minutes on the timeline, and the Analyze Audio button
3

Verify the voice

In Verify Voice Ownership, choose the prompt language, click Record, and read the text on the screen aloud (up to 30 seconds). Click Stop. Fish Audio sends the recording at once and compares it with your uploaded clips.If verification fails, you see a reason such as “The words you read didn’t match the prompt.” or “Your voice didn’t match the uploaded audio.” Try again in a quiet room and read the prompt exactly. You have a limited number of attempts.
The Verify Voice Ownership dialog with the Myself and Proxy tabs, a language selector, the prompt text, and a Record button
If the voice owner cannot record, open the Proxy tab instead. Upload proof that you are authorized (PDF, JPG, or PNG, up to 10 MB per file), accept the declaration, and click Submit. A person on the Fish Audio team reviews it.
4

Wait for training

After verification, training starts and you see Preparing training materials… Training usually takes 10 to 30 minutes, and Fish Audio sends you an email when it is done. The page updates itself, so you can leave and come back.
The training status showing Preparing training materials, with the model ID, language, training time, and state
5

Preview the voice

When training is done, the Preview section has ready-made samples. Click play to listen. To hear your own words, click Custom preview text, type a sentence, and click Generate Preview.
The Training complete banner and the Preview section with three samples and the Custom preview text button
6

Add voice details

Click Continue to Release. On the Voice Details form, Fish Audio fills in some fields from the analysis. Enter a Voice Name (required) and check the other details: Cover, Description, Gender, Age, Use cases, and Voice Style.Tick the confirmation box (“I confirm I have permission to create and share this voice model.”) and click Create Voice. Fish Audio saves the details, opens the voice page, and shows the Release Voice dialog.
The Voice Details form for a professional voice, with the name, description, tags, type, and the confirmation box
7

Release the voice

In the Release Voice dialog:
  • Visibility: choose Public or Unlisted.
  • Notice Period: choose 1 year, 2 years (recommended), or 3 years. After release, the voice stays available to the people who use it until the notice period ends. You can extend the notice period later, but you cannot shorten it.
  • Payouts: your earnings go to your team wallet. To withdraw them, the team owner connects Stripe in Team settings → Wallet. You can do this at any time after you release the voice.
Read the list of settings that you cannot change after release (training materials, voice owner, billing method, visibility, and Platform/API availability). You also cannot shorten the notice period, but you can extend it. Tick the confirmation box and click Release Voice.
The Release Voice dialog with Earn from your voice, the locked settings, Visibility, Notice Period, Payouts, and the confirmation box

​
After release

On the voice page, open the More actions (…) menu and click Edit Release to see the Release status. From there you can:
  • Extend notice period.
  • Retire voice: the voice stays available for existing Platform and API use until the notice period ends.

​
Resume a clone in progress

The Create Voice hub lists your clones below the method cards:
  • Professional Clone Draft → Continue: resume a clone that is still uploading, analyzing, or waiting for verification. You can also delete a draft.
  • Professional Clone Training in progress: the clone is verifying or training. The row updates automatically.
  • Professional Clone Awaiting Release → Release: training is done and the voice is ready to release.
  • Professional Clone Trained → Release or View: release the voice, or open it if it is already released.
  • Professional Clone Failed → Contact Support: verification or training did not succeed. You can delete a clone that failed verification.

​
Troubleshooting

PVC needs a paid plan. Upgrade your plan to create studio-quality voice clones.
You have used all your PVC slots. Upgrade your plan to get more slots.
Usually the cause is background noise, no clear speech, or more than one speaker. Record again in a quiet room with one speaker and upload the new clip.
The clone used all of its verification attempts and cannot continue. Delete it from the Create Voice hub to free the slot, or contact support.