How to make a Reel in another language in your own cloned voice

Last updated 2026-08-25 · 7 min read

What is the short version?

Reaching another market usually means either subtitles or somebody else's voice. Cloning yours and generating the script in the target language keeps the video recognisably you, which matters when the brand is a person rather than a company.

1. Go to Vizard Agent and supply a sample of your voice.
2. Give the script written in the target language.
3. Say the exact duration if the placement requires one.

What do you need before you start?

A voice sample and a translated script. Vizard Agent handles the cloning, the presenter and the graphics, so the two inputs that must be yours are the voice and the wording. Writing the script in the target language rather than asking for a translation gives you control over the register.

• A voice sample. Clean, of you speaking.
• The script. Written in the target language, in the register you want.
• A reference image. For the presenter, if one should appear.
• The exact length. Some placements are strict about it.
• What the cards should show. The steps, the offer, the features.

What do you type into Vizard Agent?

State the language and the duration as hard requirements. Vizard Agent treats an exact length as something to hit by adjusting the read's tempo rather than by trimming the ending, which is what lets a script land precisely without sounding cut off.

Prompt

Variants worth knowing:

• A regional variety. Say which one; the difference is audible to native speakers.
• On-screen demo cards. Rather than the presenter talking for the whole clip.
• A three-phase structure. Presenter, demonstration, then both together.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real cloned-voice Reel. The tempo adjustment is the technical centre of it, and the card work at the end took three passes to get the widths and spacing right.

1. Inspects the supplied files and reviews the character reference image.
2. Generates the presenter's scene from that reference and reviews it.
3. Lip-syncs the avatar to the audio and verifies the resolution and duration.
4. Extracts a frame to check the synchronisation before continuing.
5. Designs the on-screen demonstration cards and checks them.
6. Regenerates the cards with clean vector iconography, then adjusts the dimensions and spacing.
7. Corrects a badge width for a professional finish and re-checks that card.
8. Clones the voice with the script in the target language and checks the resulting duration.
9. Adjusts the cloned audio's tempo to land on exactly eighteen seconds, regenerates the facial animation to the corrected audio, builds the captions, measures the mix, then refines the card composition twice more.

Step 9's tempo adjustment is what makes an exact duration possible. Trimming to length cuts a word; slowing or tightening the read by a few percent lands the same script on the same second with nothing lost.

What does the result look like?

From the run this page is written from, probed on the delivered file: 1080x1920, H.264, 24fps, 17.96 seconds, AAC audio. Vertical, eighteen seconds to the frame, your cloned voice speaking the target language, a lip-synced presenter, vector demonstration cards and animated captions.

Landing within a frame of eighteen seconds is the point of the tempo work. A placement that specifies a duration usually enforces it, and a video that runs over gets truncated at exactly the moment the call to action appears.

When does this not work well?

Cloning a voice and putting it into a language you may not actually speak is a powerful thing to do and it has some obvious failure modes. Vizard Agent generates the read faithfully from what you write, and everything to do with accuracy and permission stays firmly with you.

• Clone only your own voice. Or one you have explicit written permission to use.
• Have the script checked by a native speaker. Vizard Agent reads what you wrote, including the errors.
• A clone in a language you do not speak is risky. You cannot hear whether it sounds right.
• Regional varieties differ audibly. Say which one you want.
• Do not imply you are fluent. The video will suggest it whether you intend that or not.

How do you fix a result that came back wrong?

Name the line or the card that is wrong. Vizard Agent keeps the cloned audio, the tempo-adjusted version of it, the presenter animation and every card revision, so a rewritten line or a resized badge is a re-render rather than a rebuild of the Reel.

• "A native speaker says that phrase is wrong." Corrected, re-cloned and re-synced.
• "The card spacing looks amateur." Rebuilt, as this run did three times.
• "It needs to be fifteen seconds." Re-timed by tempo rather than trimmed.

How does Vizard Agent compare to doing it yourself?

By hand this means hiring a voice actor in the target market, which loses the thing that made the video yours, or reading a phonetic transcription yourself, which native speakers notice immediately. Hitting an exact eighteen seconds then means either rushing the read or cutting the ending.

Common questions

Why clone rather than hire a voice in that market? Because for a personal brand the voice is the brand. A hired actor produces a competent video that no longer sounds like you, which defeats the reason your audience follows you in the first place.

Whose voice can Vizard Agent clone? Yours, or one you have explicit written permission to use. Nothing else.

Do I need to speak the language? No. Have a native speaker check the script before Vizard Agent generates it, because you will not be able to judge the result yourself.

Can Vizard Agent hit an exact duration? Yes, by adjusting the read's tempo rather than trimming the ending.

Will there be a presenter on screen? If you supply a reference image, yes, lip-synced to the cloned audio.

Why adjust tempo instead of cutting? Because a script written for an eighteen-second slot rarely generates as exactly eighteen seconds, and trimming removes a word. A small tempo change keeps every word and lands on the frame.

Can Vizard Agent do a specific regional variety? Yes, and it is worth naming, since the difference is obvious to native speakers.

Does Vizard Agent add captions? Yes, in the target language, from the generated audio's own word timings.

Can Vizard Agent build demo cards? Yes, as vector graphics rather than screenshots, and it revised these three times.

Does this work for several languages at once? Yes, as separate runs sharing the same clone and the same card designs.

Does Vizard Agent check the finished Reel? Yes. It reviews quality-control grids after each composition change before delivery.