Playground
Clone your voice
Make an Instant clone of your voice from 10 to 30 seconds of audio, confirm it is you by saying a spoken code, and use it in Text to Speech.
The Instant tab on the Voice Cloning page makes a copy of your voice from one clip of 10 to 30 seconds. You prove the voice is yours by saying a short sentence and a 4-digit code out loud. The voice is ready a few seconds later and you can use it in Text to Speech. For a closer match from longer audio, see Train a professional voice.
| Instant or Professional | Instant | Professional |
|---|---|---|
| Audio you need | 10 to 30 seconds | 10 minutes to 3 hours |
| Ready in | A few seconds | About 30 minutes, an estimate |
| Price | $2.00 fee, 500 credits to build, plus the checks, and transcription and cleaning of your sample | $5.00 fee, 1,800 credits per training hour, plus 1,100 credits per minute of your audio for transcription and cleaning |
The $2.00 fee is about 57,143 credits at the extra-credit rate. The $5.00 fee is about 142,857. See Train a professional voice and Voice Cloning limits and credits.
What you need
- A workspace with credits. Cloning shows its price before you confirm. The charges are on Voice Cloning limits and credits.
- The
clone.usepermission. The default Owner, Admin and Member roles have it. - A microphone the browser may use, or one audio file of your voice.
- A quiet room. See Record a good sample for cloning.
Steps
Step 1: Open Voice Cloning and pick a dialect
Open Voice Cloning and stay on the Instant tab. Pick the Dialect of the voice you are about to record. Emirati is the default. The list shows a dialect only once it has live voices.
Step 2: Record or upload one clip
Press Record and read the line shown under Read this out loud, or choose Upload or drag a file and pick a file. Keep the clip between 10 and 30 seconds. A recording stops by itself at 30 seconds.
- The page shows how long the clip is, and reads Good length when it is long enough.
- You can use any audio file the browser can open. The page converts it first.
- There is one clip only. A new clip replaces the old one.
- Press Play to check the clip, or Remove to start again.
Step 3: Press Clone voice
When the status reads Ready, press Clone voice. Before you confirm, the page shows This voice costs about N credits., with N filled in for your workspace.
Step 4: Confirm it is you
The page opens Confirm it is you and shows Say this out loud: a sentence and a 4-digit code. Press Record, say the sentence, then say each digit of the code, and stop. The code recording is 3 to 10 seconds long.
The page explains: This keeps your voice yours: only the person who says the code can clone it. A code uses the digits 1 to 9 and runs out after 10 minutes. Press New code for another one. If the check fails the page reads Not clear, try again and gives you a new code.
Step 5: Use the voice
Your voice is ready in a few seconds. A message pops up: Your voice is ready. The result screen has Play, Use in Text to Speech and Clone another. Type a line under Type something to hear and press Play to try it.
The voice is named My voice, then My voice 2, and so on. It is also listed under Your voices on the same page.
Get it right
- Speak naturally, in the dialect you picked, in one steady voice.
- Use a quiet room. Background noise and echo are copied too.
- Keep one speaker only, and aim for the full 30 seconds.
- Keep the microphone at the same distance for the whole clip.
- Say the sentence and the code clearly, one digit at a time.
More advice is on Record a good sample for cloning.
Worked example
Layla records 20 seconds in an Emirati voice with the Emirati dialect selected. The page reads Good length, so she presses Clone voice and reads the price shown. It includes the $2.00 fee and every line. Then she says the sentence and the code. A few seconds later the message Your voice is ready. pops up. She presses Use in Text to Speech and writes her first line with her own voice.
What happens next
The voice appears in these places:
- Your voices on the Voice Cloning page.
- The Voices page, under Recorded by you. See the Voice Library.
- The Cloned filter of the voice picker in Text to Speech. See Generate speech.
- Assets, on its Voice Cloning tab.
- EchoReader.
A cloned voice does not work in agents yet. For a closer match from longer audio, see Train a professional voice.
Rename or delete the voice
Open the voice from Your voices to rename or delete it.
- Renaming: a name has up to 60 characters. Names are unique in the workspace, and capital letters do not make a name different.
- Deleting: press Delete. The voice is deleted right away. Anything already made with it stays where it is. The consent recording is kept 12 months. Deleting frees the slot and the name at once.
Limits
| Limit | Value |
|---|---|
| Length of the clip | 10 to 30 seconds |
| Clips | One. A new clip replaces the old one |
| Code | 4 digits, expires after 10 minutes |
| Code recording | 3 to 10 seconds |
| Voices per workspace | 20. A professional voice in progress counts as one |
| Codes per user | 20 per 10 minutes |
| Code checks per user | 10 per 10 minutes |
| Name | 60 characters, unique in the workspace |
Dialects for Instant cloning: Emirati, Saudi, Egyptian, Kuwaiti, Fusha, American, British and English. Iraqi has no cloning.
Troubleshooting
Too long, keep it under 30 seconds
The clip is over 30 seconds. A long clip is refused, not trimmed. Record again, or upload a shorter file.
Add N more seconds
The clip is shorter than 10 seconds. Keep reading until the page reads Good length.
This file could not be read. Try a WAV, MP3 or M4A file.
The browser could not open the file. Export it again as WAV, MP3 or M4A and upload it.
We could not use your microphone. Check the browser permission and try again.
The browser did not allow the microphone. Allow it in the browser settings for this site and press Record again.
Not clear, try again
The code recording did not match. Say the sentence and each digit of the new code. The code is spent after each try.
Code expired, get a new one.
The code is older than 10 minutes, or it was already used. Press New code and say the new one.
A workspace can keep up to 20 cloned voices. Delete one first.
The workspace is full. Delete a voice you no longer need, then press Clone voice again.
A voice with this name already exists. Choose another name.
Another voice in the workspace has this name, in any capital letters. Pick another name.
Price not set yet. Voice cloning opens once its price is published.
Cloning has no price yet, so it is closed. Nothing was charged. Try again later.
The two recordings are not the same person.
The clip and the code recording do not sound like one speaker. Record both yourself, in the same room, and try again.
We could not confirm it is you. Try again.
The code recording did not pass the check. Say the sentence and each digit of the new code, clearly.
Confirm it is you again.
The check was used up, or the clip changed after it. Press Clone voice and say a new code.
Too many codes. Wait a few minutes and try again.
You asked for more than 20 codes in 10 minutes. Wait, then press New code.
Too many tries. Wait a few minutes and try again.
You made more than 10 code checks in 10 minutes. Wait, then try again.
Insufficient subscription balance
The balance cannot pay the price. Add credits, then press Clone voice.
Your voice could not be made ready. You were not charged.
The build failed and the charge was refunded. Try again.
Voice cloning is busy, try again in a minute.
Too many clones are running. Wait a minute and try again.
Something went wrong. Try again.
The request failed. Try the step again. If it repeats, wait a few minutes.
FAQ
Can I use a cloned voice in an agent?
Not yet. The error reads: "Cloned voices cannot be used in agents yet." Use it in Text to Speech, on the Voices page under Recorded by you, in Assets on its Voice Cloning tab or in EchoReader.
Can I clone someone else's voice?
No. You must say the sentence and the code yourself, and only the person who says the code can clone the voice.
How much does it cost?
The shown price includes every part: the $2.00 fee, 500 credits for the build, the checks, the transcription and cleaning of your sample, and the preview line. The fee is about 57,143 credits at the extra-credit rate and 63,158 at the Starter rate. Other plans have other rates, on Credits. See Voice Cloning limits and credits. Using a cloned voice in Text to Speech costs the same as any voice.
Can I upload more than one clip?
No. Instant cloning takes one clip. For more audio, see Train a professional voice.
Who can clone?
Anyone with the clone.use permission. The Owner, Admin and Member roles have it.
Related
- Record a good sample for cloning: how to get a clean recording.
- Train a professional voice: a closer match from 10 minutes to 3 hours of audio.
- Voice Cloning limits and credits: every limit and charge.
- Voice Library: where your cloned voice is listed.
- Generate speech: use the voice in Text to Speech.