Vizard Agent

How to remake one of your videos with a new voice and new visuals

Last updated 2026-08-16 · 6 min read

Upload the video you want to remake to Vizard Agent. Vizard Agent transcribes it to recover the script, rewrites that script, generates a new voiceover, sources fresh stock footage for every section, and times the visuals and captions against its own narration rather than the original.

What is the short version?

Your best video was probably posted once, on one platform, in one shape, a year ago. Remaking it means recovering the script from the video itself, giving it a new read and new pictures, and posting it as a new thing rather than a repost.

  1. Go to Vizard Agent and upload the original video.
  2. Say what should carry over — the message and the structure, not the exact wording.
  3. Say the voice, the shape and whether you want captions.

What do you need before you start?

The original video and the right to remake it. Vizard Agent recovers the script and rebuilds everything around it, so what it needs from you is a decision about how much should change: a light refresh and a full rewrite are different requests with different outcomes.

What do you type into Vizard Agent?

Say what to preserve and what to replace. Vizard Agent reads the original itself, so you do not need to retype the script, and one sentence about what the video is for shapes the rewrite more than any amount of description.

Prompt

Recreate this video with a rewritten script, a new voiceover and new visuals. Keep the message and the structure. Vertical, captions on.

Variants worth knowing:

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real remake. The important detail comes near the end: the new visuals are timed against the new narration's own word timings rather than inherited from the original video's pacing, which is what stops the footage drifting away from the script.

  1. Transcribes the original video to recover its script.
  2. Probes the file and extracts frames, tiling them to see the layout and style.
  3. Looks at the tiled frames to understand how the original presented itself.
  4. Loads the explainer template and playbook rather than improvising a structure.
  5. Checks the voiceover options and pricing, then auditions a narrator.
  6. Generates the new voiceover and checks its duration.
  7. Transcribes its own new narration to get precise word timings.
  8. Searches for portrait-oriented stock footage for every section of the script.
  9. Downloads the selected clips in parallel and finds a background track to match the tone.

Step 7 is the habit worth copying. Vizard Agent measured its own generated narration rather than trusting the requested length, which is what keeps the footage changing on the sentence rather than drifting a second further out with every section.

What does the result look like?

From the run this page is written from, probed on the delivered file: 1080x1920, H.264, 30fps, 43.6 seconds, AAC audio. Vertical, forty-four seconds, narrated by a new voice over freshly sourced footage, with captions, delivered as one file from an original that was neither vertical nor silent.

Ask for captions after the fact and Vizard Agent adds them from the timings it already has, which is a fast change rather than a rebuild.

Forty seconds is a natural length for a remake. The original was longer, and a rewrite is the moment to cut the parts that were only there because they were filmed.

This category was not separately measured, so treat the timing as a range rather than a promise: comparable work runs a median of 28 to 38 minutes end to end. Across all projects the median cost by tier is Flash 47, Pro 55, Max 242, Ultra 263 credits.

When does this not work well?

This is a genuinely useful technique on your own material and a bad idea on anyone else's. Rewriting a script and changing the voice does not make somebody else's video yours, and Vizard Agent has no way to know whose video you uploaded.

How do you fix a result that came back wrong?

Say what to change. Vizard Agent keeps the recovered script, the new narration, its word timings and the downloaded footage, so a rewritten line re-reads one section, and a footage swap leaves the narration and its timing untouched.

How does Vizard Agent compare to doing it yourself?

By hand this is transcribing your old video, rewriting it in a document, recording a new read, hunting stock footage section by section, and timing it all. The transcription and the timing are the mechanical parts, and they are most of the hours.

By hand Vizard Agent
Recovering the script Transcribe it yourself Transcribed automatically
The new read Record and re-record Cast and generated
Sourcing visuals Search per section Searched and downloaded in parallel
Timing Nudge each clip From the new narration's word timings

Common questions

Can I do this with someone else's video? No. Changing the voice and visuals does not make the underlying work yours, and Vizard Agent is built for remaking your own material.

Will the new script say the same thing? It keeps the message and the structure. Read it — Vizard Agent can rewrite any part you want closer to the original.

Can it use my voice? Yes. Provide a clean sample and Vizard Agent clones your voice for the new read.

Can it use my own footage instead of stock? Yes, and it should. Upload what you have and Vizard Agent prefers your material.

Can it change the aspect ratio? Yes. This is a good moment to move a wide video to vertical, since everything is being rebuilt anyway.

Will the captions match the new voice? Yes. They come from word timings measured on the generated narration itself.

How different will it look from the original? Completely, unless you ask otherwise. New voice, new footage, new pacing — only the argument carries over.

Can it remake several old videos at once? Ask one at a time. Each needs its own transcription and its own footage search.