Playground

Voice Cloning limits and credits

The limits on Instant and Professional voice cloning, what a clone costs in credits, when you are charged, and what happens when the balance cannot cover it.

This page is the lookup table for Voice Cloning: what the Instant and Professional tabs accept, what a clone costs in credits, when the charge is taken, and what you see when the workspace balance cannot cover it. What a credit is and the rates are on Credits.

Limits

Limit Instant Professional
Audio One clip of 10 to 30 seconds 10 minutes to 3 hours in total
Files One. A new clip replaces the old one Up to 20, 2 GB in total
Clear speech Not a separate limit At least 20 minutes after cleanup
Recording in the browser Stops at 30 seconds Stops at 30 minutes
Code recording 3 to 10 seconds 3 to 10 seconds
In progress Ready in a few seconds One professional voice per workspace
Voices per workspace 20. A professional voice in progress counts as one Same
Name Up to 60 characters, unique in the workspace Same
Codes per user 20 per 10 minutes Same
Code checks per user 10 per 10 minutes Same
Who can clone Anyone with the clone.use permission Same

What a clone costs

Clone Fee Build or training
Instant $2.00 500 credits for the build
Professional $5.00 1,800 credits per training hour

The dollar fee is taken in credits at the rate of the credits that pay. On the rates in Credits, the fee is:

Credits that pay Instant, $2.00 Professional, $5.00
Extra credits (28,571.43 per $1) 57,143 142,857
Starter monthly and trial (31,578.95 per $1) 63,158 157,895

Other plans have other rates. The fee is rounded to a whole credit on its own.

Every clone also pays for its checks, its transcription, its cleaning and its preview line:

Charge Rate
Liveness check 2 credits per second of the code recording
Same-speaker check 1 credit per second compared
Speech to Text 100 credits per minute
Sound cleaning 1,000 credits per minute
Preview line 1 credit per character

The checks, the transcription and the cleaning are charged per second. The credit lines of one clone, preview line included, are added up and rounded once to a whole credit.

When you are charged

Moment What happens
Instant, when you confirm The whole price is charged before the voice is built, and refunded in full if the build fails.
Professional, at the start The price is charged for a 1 hour training estimate, fee included.
Professional, training ends Only the difference in training time is charged or refunded.
A clone fails The charge is refunded in full.
You use the voice Text to Speech costs the same as with any voice.

When the balance cannot cover a clone

The page shows an estimate before you confirm: This voice costs about N credits. For Professional it reads This voice costs about N credits with 1 hour of training. The final price is set when training ends.

If the workspace balance is lower than the estimate, the request is refused with Insufficient subscription balance. Add credits before you confirm. The ways to add credits are on Credits, top-ups and rollover.

If cloning has no published price, the page reads Price not set yet. Voice cloning opens once its price is published. Nothing is charged. Try again later.

Using a cloned voice

A cloned voice works in Text to Speech (the voice picker has a Cloned filter), on the Voices page under Recorded by you, in Assets on its Voice Cloning tab and in EchoReader. See Generate speech and Voice Library. It costs the same as any voice. It does not work in agents yet.

Worked example

An Instant clone from a 20 second sample with a 10 second code recording:

Line Quantity and rate Credits
Speech to Text, code 10 seconds at 100 per minute 16.67
Liveness check 10 seconds × 2 20
Same-speaker check 10 + 20 = 30 seconds × 1 30
Sound cleaning, sample 20 seconds at 1,000 per minute 333.33
Speech to Text, sample 20 seconds at 100 per minute 33.33
Build 1 voice 500
Preview line 1 per character of the line P
Fee $2.00, taken in credits F

The lines above P and F add up to 933.33 credits. Those lines and the P preview characters are added and rounded once to a whole credit. The fee is converted to credits and rounded on its own: F is 57,143 at the extra-credit rate, so the total is about 58,076 + P credits. The page shows the total before you confirm: This voice costs about N credits.

A Professional clone from 45 minutes of audio with a 10 second code recording:

Line Quantity and rate Credits
Speech to Text, code 10 seconds at 100 per minute 16.67
Liveness check 10 seconds × 2 20
Same-speaker check 10 + 30 = 40 seconds × 1 40
Speech to Text, audio 45 minutes at 100 per minute 4,500
Sound cleaning, audio 45 minutes at 1,000 per minute 45,000
Training 1 hour estimate at 1,800 per hour 1,800
Preview line 1 per character of the line P
Fee $5.00, taken in credits F

The lines above P and F add up to 51,376.67 credits. With the P preview characters they are rounded once. F is 142,857 at the extra-credit rate, so the total is about 194,234 + P credits. That is the 1 hour estimate. When training ends, only the difference in training time is charged or refunded. The audio alone costs 1,100 credits per minute.

FAQ

Can I see the price before I confirm?

Yes. The page shows This voice costs about N credits. before the code step. See Clone your voice.

Am I charged if a clone fails?

No. A failed clone is refunded in full.

Does a longer sample cost more?

Yes. For Instant, the transcription, the cleaning and the same-speaker check are charged per second, so a longer sample adds to the total. For Professional, the same-speaker check compares the code recording with 30 seconds, so it does not grow. The transcription and the cleaning of all your audio do: 1,100 credits per minute.