Rewrite a Talking Video with a New Script

This Text to Video Lip Sync example starts with the same natural talking clip as the uploaded-audio tutorial, but uses a different input strategy. Instead of supplying the final line as audio, we clone a voice from a short reference sample and replace the spoken line with text.

Source clip

Here is the source video used for this tutorial.

Source video requirements

For text-driven rewrites, the source clip should still feel like a natural talking performance.

• A natural speaking clip works best, and it does not need to contain the final line you want in the finished video
• Around 5 seconds to a few minutes is fine
• The face should stay visible for most of the shot
• Small, believable motion is good
• Avoid exaggerated gestures, sudden turns, or large action beats

If you are filming a clip specifically for this workflow, keep it simple and natural. Saying "one two three" during recording is enough. That is one of the main strengths of this tool: you do not need to memorize or perform the final script on camera to get a believable talking video later.

Voice reference

This is the voice reference used to clone the speaking style:

Replacement script

This is the exact script used for the rewritten version:

That's it for today! Isn't this incredible? It's free, it's fast, it's u* — and you can even generate lip sync videos up to 60 minutes long.

This path is useful when:

• You want to revise the line without recording new audio
• You need to localize or patch one sentence quickly
• You want the new speech to follow a chosen voice identity

Generated result

Here is the finished result generated from the source clip, voice reference, and replacement script:

Open the dedicated watch page for this result

What this tutorial shows

• One natural source clip can be reused for a fully new spoken line
• A short voice reference is enough to define the vocal identity
• Text rewrites are strongest when the source motion stays calm and believable

How to recreate this workflow

1. Open Text to Video Lip Sync.
2. Upload a natural talking video with one clear face.
3. Upload a short voice reference for cloning.
4. Paste the new script you want the video to say.
5. Generate the result and check whether the new spoken line feels believable in the original shot.