How to fit a long text into a scrolling video at a readable speed

Last updated 2026-09-10 · 8 min read

What is the short version?

A scrolling text video has exactly one real constraint, which is reading speed. Pick the length first and the text either races past unread or runs out halfway. Pick the speed first and hand it to Vizard Agent, and the length is simply however long the words take.

1. Go to Vizard Agent with the text.
2. Say the whole thing must fit, nothing cut.
3. Say the scroll should be comfortable to read.

What do you need before you start?

The text, in whatever form you have it. If it only exists as pictures — a screenshot, a slide, frames of another video — that is fine and worth mentioning, because it has to be read off the images before anything can be built from it.

• The text. Typed, or in pictures.
• That all of it must appear. Say so explicitly.
• The reading speed. "Comfortable" is a real answer.
• What is behind it. Photographs, a colour, footage.
• Any music. And whether it sets the length.

What do you type into Vizard Agent?

Say the speed and let the length follow. Starting with "make it thirty seconds" fixes the wrong variable, and every round after that is a negotiation between the runtime you named and the text you actually want on screen.

Prompt

Variants worth knowing:

• "Fit all of the text." Nothing gets trimmed to fit a length.
• "Comfortable to read." The constraint that matters.
• "As long as it needs to be." Lets the runtime follow.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real scrolling-text video whose source text existed only as frames of another video. The first third is reading; the last third is checking the roll in more than one place.

1. Reviews the full sequence of text frames in the source.
2. Prepares full-resolution checkpoints so the text can be read exactly.
3. Reads each frame in turn — opening, middle, closing.
4. Loads the right approach for a caption roll over stills.
5. Generates background music to suit the piece.
6. Reviews every uploaded photo before using any of them.
7. Inspects the options for a text roll specifically.
8. Builds the vertical background and the rolling caption track at the full length.
9. Renders a preview and reviews it across its whole duration.
10. Inspects readability at several points in the finished render.
11. Checks the roll late, in the middle and at the final frame.
12. Fixes the scrolling and re-renders.
13. Verifies the corrected roll from start to finish.

Step two is what makes the text trustworthy. Reading words off a video frame at preview size produces plausible mistakes, and a scrolling piece is nothing but text, so the reading is done on full-resolution checkpoints rather than on thumbnails.

Steps ten and eleven are the check people do once. A roll can be comfortable at the start and crowded by the end — the speed is constant but the line lengths are not — so readability gets sampled at several points across the whole run.

What does the result look like?

Every word of the text on screen, moving at a pace you can read without leaning in, over the pictures you supplied, for however long that turns out to take. Nothing was cut to make it fit a number.

If two minutes is what the text needs, the video is two minutes.

When does this not work well?

Reading speed and attention pull in opposite directions. A text long enough to need five minutes will be read by very few people, and no scroll speed fixes that — the honest options are cutting the text or splitting it into parts.

• Very long texts. Length becomes the problem.
• A hard runtime. Then something has to be cut.
• Dense scripts at small sizes. Legibility caps the speed.
• Music with a fixed length. It will not match the reading.
• Text over busy footage. Contrast limits the size.

How do you fix a result that came back wrong?

Say whether it is too fast, too slow or simply unreadable, and roughly where along the roll. Vizard Agent keeps the text it read and the timing of the roll it built, so a speed change is a single adjustment rather than a rebuild of the whole piece.

• "Too fast at the end." The speed evened across the roll.
• "I cannot read it over the photo." Contrast raised behind the text.
• "Some text is missing." Re-read from the source and reinserted.
• "Make it longer." The roll slowed rather than the text cut.

How does Vizard Agent compare to doing it yourself?

By hand the length gets chosen first, because that is what the brief usually says, and the text is then squeezed or padded to match. Readability gets judged on the first few lines, which are always the most comfortable ones.

Common questions

Can it read text out of a video? Yes. Vizard Agent reads it at full resolution rather than guessing.

Will all my text fit? Yes, if you let the length follow. Vizard Agent will say what it needs.

What speed is comfortable? Vizard Agent sets one and shows you; adjust from there.

Can I set a hard runtime? Yes, but then some text has to go, and Vizard Agent will say which.

Can it use my photos as the background? Yes. Vizard Agent reviews each one before using it.

Will the text be readable over them? Vizard Agent checks, and raises the contrast behind it where needed.

Can it add music? Yes, written to the final length rather than looped to it.

Does it work vertically? Yes. Vizard Agent finds vertical is the usual case for this.

Can it read the text aloud too? Yes. Ask Vizard Agent for narration alongside the roll.

What about another script or language? Yes, though Vizard Agent has to find a typeface that supports it.

Can I split it into parts? Yes. Vizard Agent will break it at sensible points.

Will the speed be constant? Yes unless you ask otherwise; Vizard Agent checks it holds throughout.

Can I see it before the full render? Yes. Ask Vizard Agent for a preview of the roll.

Why not just pick a length? Because the length is a consequence of how much text there is, and choosing it first means choosing to lose some of the text without deciding which.