Can I use the audio outside Backlot?

Yes, that's the point of the download. Takes come back as standard MP3 files you can drop into Premiere, Resolve, CapCut, a podcast edit, an e-learning module or a phone system. Nothing about the file ties it back to the app.

Do I have to make a video to use a voice?

No. Voices are a service in their own right. Clone or design one, paste a script, download the audio, and never open a production if you don't need one.

How long does cloning take, and how good is it?

Processing takes about a minute. The quality follows the recording: thirty seconds gives a rough match, a minute gives a good clone, two minutes gives the closest match. Read at your normal pace in a quiet room and the difference is obvious.

Whose voice am I allowed to clone?

Your own, or one you have written permission to use. That's a condition of the service, not a formality: cloning a person's voice without their consent is prohibited and gets the voice removed.

Cloning a voice is 25 credits ($3.00); designing one from a description is 15; picking a stock voice is free. After that you pay 1 credit per 250 characters of script, roughly 4 credits ($0.48) for a spoken minute.

Can a voice speak another language?

Yes. Write the line in the language you want and the same voice delivers it, keeping its own timbre. In a production the lip sync follows the audio, so the mouth matches whatever language was spoken.

Voices

A voice of your own.Yours to take anywhere.

Clone your voice by reading a script into your browser, design one from a description, or pick from the library. Then have it read anything you like and download the file, with or without ever making a video.

Make a voice See what it costs

Open one to hear it. Every line was spoken by a voice made in Backlot.

Three ways to get a voice.

However you make it, it behaves the same afterwards: auditionable, castable, and downloadable.

Clone your own

Read a script into your microphone on the page: no recording app, no file to find. Three lengths, from a thirty-second rough match to two minutes for the closest clone we can build. 25 credits, once.

Design one from words

Describe the voice you want (“warm, mid-forties, slight Irish lilt, unhurried”) and Backlot builds it. Adjust stability and similarity until the delivery sits where you want it. 15 credits, once.

Start from the library

Pick a ready-made voice and go; that one is free. Every voice, however it was made, lands on the same shelf and can be auditioned on your own line before you commit to it.

Cloning

Read it here. No recording app, no file to find.

Press record and read the script on screen. Backlot shows your level as you speak, trims the silence at either end and hands the clean take straight to the cloning model. If you'd rather use a recording you already have, upload it instead.

Three script lengths (about 30 seconds, a minute, or two) because a longer read really does clone better
A live level meter and a retake button, so you can hear whether the room is too noisy before you commit
Takes under eight seconds are rejected rather than quietly producing a bad clone

Clone your voice

Quick

≈ 30 s

Standard

≈ 1 min

Best

≈ 2 min

Hi, this is my voice, and I’m recording it so the studio can use it later. I’ll try to speak the way I normally do: not too fast, not too slow, with the pauses I’d use…

Stop0:41 / ≈1:00

Voiceover

Paste a script. Download the audio.

Give the voice up to 5,000 characters, about six minutes of speech, and it reads the lot. Every take is kept: name it, play it, download the MP3 and take it wherever the work actually lives. Dictate the script with the microphone if you'd rather talk than type.

Standard MP3 at 44.1 kHz: drop it into any editor, podcast, module or phone system
Takes stack up instead of overwriting each other, so you can compare two readings and keep both
1 credit per 250 characters, about 4 credits for a spoken minute, and nothing more to pay for the voice itself

Generate a take

Script

Welcome back. Last time we left the crew halfway up the gantry, waiting on a permit that hadn’t been signed…

218 characters · about 15s spoken · 1 creditGenerate take

Episode 4 intro
Welcome back. Last time we left the crew halfway up the gantry…
ready
0:18
Download MP32 credits
Safety refresher VO
No permit, no hot work. Every time, on every job, without exception.
ready
0:09
Download MP31 credit
Store announcement
We close at six tonight. Tills shut at quarter to.
ready
0:06
Download MP31 credit

The shelf

One library, the whole team casting from it.

Cloned, designed and library voices sit together in the workspace, each with its gender, age, accent and language on the card. Audition any of them on your own line before you decide, and assign the one you want to a character so every scene they appear in speaks in that voice.

Shared across the studio, with the same roles and permissions as everything else
Assign a voice to a Troupe character once and every production inherits it
Only voices that match the character's gender are offered, each with a sample to hear first
Change a voice mid-production and the shots that character speaks in are redone in it, price on the button
Archive a voice without losing the takes you already generated from it

See the library

Malya, site leadcloned
EN · female · 30s · Levantine
Narrator, warmdesigned
EN · male · 50s · RP
Ava, brand hostcloned
EN · female · 20s · American
Announcer, brightlibrary
EN · neutral · studio

Audition any of them on any line before you commit.

And when you do want a video, it's already cast.

The same voice that read your script can perform it on camera.

See how video works

Performed, not pasted on

In a production the dialogue is generated first, then the shot is rendered from that audio, so mouth shapes, timing and gesture come from the line itself.

Any language, same person

Write the line in another language and the same voice delivers it. Lip sync follows the audio, so a dubbed cut doesn't look dubbed.

Consent, on the record

Clone your own voice or one you have written permission to use. Every voice records who created it and when, and can be removed on request.

One price to build it. Then you pay for what it says.

Building a voice is a one-off, and it stays yours. After that, speech is priced by the character. Credits are bought outright at a fixed price, so nothing expires at the end of the month.

Make a voice Full pricing

Cloning your voice: 25 credits
Designing one from a description: 15 credits
Taking one from the library: Free
250 characters of speech: 1 credit
A spoken minute, roughly: 4 credits

Questions about voices

Something we haven't answered? Talk to us.

Can I use the audio outside Backlot?: Yes, that's the point of the download. Takes come back as standard MP3 files you can drop into Premiere, Resolve, CapCut, a podcast edit, an e-learning module or a phone system. Nothing about the file ties it back to the app.
Do I have to make a video to use a voice?: No. Voices are a service in their own right. Clone or design one, paste a script, download the audio, and never open a production if you don't need one.
How long does cloning take, and how good is it?: Processing takes about a minute. The quality follows the recording: thirty seconds gives a rough match, a minute gives a good clone, two minutes gives the closest match. Read at your normal pace in a quiet room and the difference is obvious.
Whose voice am I allowed to clone?: Your own, or one you have written permission to use. That's a condition of the service, not a formality: cloning a person's voice without their consent is prohibited and gets the voice removed.
What does it cost?: Cloning a voice is 25 credits ($3.00); designing one from a description is 15; picking a stock voice is free. After that you pay 1 credit per 250 characters of script, roughly 4 credits ($0.48) for a spoken minute.
Can a voice speak another language?: Yes. Write the line in the language you want and the same voice delivers it, keeping its own timbre. In a production the lip sync follows the audio, so the mouth matches whatever language was spoken.

Hear yourself read it back.

Create a studio, record a minute of script and generate your first take. You'll know straight away whether the clone is good enough for the work you have in mind.

Make a voice Talk to us