How to turn a script into a video
When the script already exists, the job changes from writing to preparing. Most scripts written for a person to read need adjustment before a generator handles them well.
The adjustments are small and specific, and making them before you generate saves regenerating afterwards.
This produces a finished vertical video from your existing words.
Steps
Check the length against the runtime
Roughly 2.5 words per second. A 60-second video is about 150 words, so a script written without a target is usually two or three times too long. Cut before generating rather than discovering it in the output.
Remove anything that only works on a page
Parentheticals, footnotes, headings and anything relying on the reader looking back. Spoken delivery moves in one direction with no ability to re-read, so a clause that clarifies in print confuses aloud.
Shorten the sentences
Written prose tolerates long sentences; narration does not. If a sentence needs a breath in the middle, split it. This is the change that most affects whether narration sounds like telling rather than reading.
Generate the video
Paste the prepared script into Leaxor and pick a tier. Scenes are generated per beat, narration is timed to them, and captions are burned in.
Review before regenerating
If the result feels off, the cause is almost always in the script rather than the generation. Regenerating an unchanged script costs credits and produces a comparable result, so fix the words first.
What you end up with
A vertical MP4 whose scene count follows the script's beats, with narration timed to the animation and captions carrying the words for muted playback.
Watch sample videosWhen this isn't the right approach
If the script depends on a specific performance, comic timing or an accent, generated narration will flatten it. Record it yourself and use an editor. If it is a dialogue between characters with distinct voices, that also sits outside what this pipeline does well.
Ready to make one?
Turn a script into a videoFrequently asked
Can I control which scene goes where?
Scenes follow the script's beats, so the practical control is the script itself. Splitting a paragraph into two shorter ones usually produces two scenes where there was one, which is the most reliable way to influence the pacing.
What if my script is much longer than 60 seconds?
Either split it into a series, which usually performs better anyway, or produce it as long-form. A 400-word script is a three-minute video, and three minutes is a different format with different pacing rather than a longer Short.
Can I reuse a script written for a presenter?
Usually with edits. Scripts written to be performed often lean on delivery, pauses and emphasis that generated narration will not reproduce. Anything whose meaning depends on how a line is said needs rewriting so the meaning is in the words instead.
Does script formatting affect the output?
Paragraph breaks do, because scenes follow beats. Stage directions, speaker labels and bracketed notes should be removed rather than left in, since they are read as content rather than instruction.
What length should I aim for?
Around 150 words for a 60-second video. Scripts arriving longer than that are not wrong, they are simply a different format, and deciding whether you want a Short or a three-minute piece before generating saves a regeneration.
Why do scripts that read well often fail on video?
Because reading lets you re-read and listening does not. Long sentences with subordinate clauses are comfortable on a page and unfollowable in audio, where a listener who loses the thread cannot go back three words. The same applies to any construction that depends on punctuation the ear cannot hear. Reading a draft aloud exposes almost all of it within a minute, and it is the single most useful check between a finished script and a finished video. Anything you stumble over reading is something a listener will stumble over hearing, so mark it and rewrite it before you generate.