Voice-over Runs on our GPU
Your script, read like a pro. Narration in five narrator styles, or in your own voice with your consent.
1 credit · under a minute on our GPU with Narrator style · English
Paste a script and pick a narrator style, or upload 5-15 seconds of your own voice. Chatterbox reads it with natural pacing and adjustable energy. Every clip carries an inaudible watermark.
What you can do
Each example is a real run on our GPU: what went in, the exact instruction, and the result exactly as the model made it. Pick one to see it, then start from the same instruction.
Only the instruction: no file.
What we asked
Every great story starts with a single step. Beyond the mountains, a quiet road waits in the dark, and only the brave will follow it. Tonight, the journey begins, and nothing will ever be the same again.
What came out, untouched
47 s on our RTX 4060 1 credit ($0.10) · Narrator style · English
Only the instruction: no file.
What we asked
And they're off! The runners burst from the line, the crowd is on its feet, and the pace is incredible. Into the final bend now, shoulder to shoulder, one last push for the finish. What a race this is!
What came out, untouched
8 s on our RTX 4060 1 credit ($0.10) · Narrator style · English
Only the instruction: no file.
What we asked
At first light, the river valley comes alive. Herons wade through the shallows, and the old forest fills with birdsong. Few places on Earth change so much in a single morning, or reward a patient observer so well.
What came out, untouched
9 s on our RTX 4060 1 credit ($0.10) · Narrator style · English
Only the instruction: no file.
What we asked
The house stood at the end of a long lane, where the hedges grew tall and the gate never quite closed. Clara set down her suitcase, took a slow breath, and stepped inside for the first time in twenty years.
What came out, untouched
11 s on our RTX 4060 1 credit ($0.10) · Narrator style · English
Only the instruction: no file.
What we asked
Big news, everyone! Our summer collection is finally here, packed with fresh colours, easy styles and something for every day of the week. Come and see it this weekend, and bring a friend along for the fun!
What came out, untouched
9 s on our RTX 4060 1 credit ($0.10) · Narrator style · English
5 examples from our own runs, with the times we measured. Open "Prompt and settings we used" under any of them for the exact input.
Deep cinematic narrator, 13 s
Whisper-small heard it with a word error rate of 0.00. Re-encoded to MP3 128 kb/s for the web; the studio delivers WAV. GPU time includes loading the model.
Prompt
Every great story starts with a single step. Beyond the mountains, a quiet road waits in the dark, and only the brave will follow it. Tonight, the journey begins, and nothing will ever be the same again.
Energetic sports commentator, 13 s
Whisper-small heard it with a word error rate of 0.00. Re-encoded to MP3 128 kb/s for the web; the studio delivers WAV.
Prompt
And they're off! The runners burst from the line, the crowd is on its feet, and the pace is incredible. Into the final bend now, shoulder to shoulder, one last push for the finish. What a race this is!
Warm documentary, 13 s
Whisper-small heard it with a word error rate of 0.00. Re-encoded to MP3 128 kb/s for the web; the studio delivers WAV.
Prompt
At first light, the river valley comes alive. Herons wade through the shallows, and the old forest fills with birdsong. Few places on Earth change so much in a single morning, or reward a patient observer so well.
Calm audiobook, 15 s
Whisper-small heard it with a word error rate of 0.03 (its spelling: '20'). Re-encoded to MP3 128 kb/s for the web; the studio delivers WAV.
Prompt
The house stood at the end of a long lane, where the hedges grew tall and the gate never quite closed. Clara set down her suitcase, took a slow breath, and stepped inside for the first time in twenty years.
Upbeat promo, 12 s
Whisper-small heard it with a word error rate of 0.03 (its spelling: 'colors'). Re-encoded to MP3 128 kb/s for the web; the studio delivers WAV.
Prompt
Big news, everyone! Our summer collection is finally here, packed with fresh colours, easy styles and something for every day of the week. Come and see it this weekend, and bring a friend along for the fun!
The prompt and the model decide the output; bigger models and settings cost more credits. These notes come from what worked, and what did not, in our runs.
Times are end to end on our GPU (an RTX 4060 Laptop GPU, 8 GB), model loading included, before any jobs ahead of yours. "Estimate" marks an option we have not measured yet.
| Option | Credits | Typical time | Slowest seen | What to expect |
|---|---|---|---|---|
| Narrator style · EnglishRecommended | 1 | 25 s | 25 s1 run | Measured on the same model and path: 15.5 s of speech in 10.9 s plus 13.9 s load. Longer scripts add about 0.7 s per second of speech. |
| Your own voice · English | 1 | 25 s | 25 s1 run | Needs 5-15 s of clean speech and the consent box. 96% of words were recognised in the English trial. |
| Your own voice · other languages | 1 | 1 minestimate | 2 min 10 s | Arabic and Hindi worked in trials (13.5-16.6 s per clip), but the model load was not cleanly measured and a native listener has not yet judged the Arabic. Time to be measured; this is a conservative estimate. |
Step-by-step recipes that use Voice-over with other studio tools, with the credits each step costs.
You answer for your prompt and your choice of model and settings; the safety filter still checks every job, and a refused prompt costs nothing.
1 creditunder a minute on our GPU, plus any jobs ahead
You'll have 7 days to download it; then it's deleted.
Results are kept 7 days. Credits come back only if the job fails or never starts.
Reading the line…
Credits are held when you send a job and spent when the result is ready. They come back automatically if the job fails, is refused or has not started within 24 hours. Results are kept 7 days, or 48 hours after your first download, then deleted; a result that expires unused is not refunded. Pricing and refunds
No. The studio only offers its own narrator styles and your own voice, uploaded with your consent.
5 to 15 seconds of you speaking clearly, alone, without music. Tick the consent box.
Yes, every clip carries an inaudible watermark so AI speech can be identified.
English. The other-languages option reads Arabic and Hindi too; those have not yet been judged by native listeners.
About 50 seconds of speech per job. Split longer scripts.