How to make a health fact video from a study you read

Last updated 2026-08-24 · 7 min read

What is the short version?

A health fact video is one finding, said clearly, with the hook on screen in the first two seconds. Everything else — the footage, the music, the captions — exists to keep someone watching long enough to hear the whole claim rather than half of it.

1. Go to Vizard Agent and state the finding in your own words.
2. Say who it is for and what they should do about it.
3. Say the length and that you want a hook card.

What do you need before you start?

The finding, stated carefully. Vizard Agent writes the script, sources the footage and builds the graphics, so the only irreplaceable input is the claim itself and how precisely you word it. In this category the wording is the product, because a finding stated loosely becomes advice you did not mean to give.

• The finding. As precisely as the study actually put it.
• The source. So it can be named or shown.
• The audience. Who this applies to, and who it does not.
• The length. Around a minute for a claim with context.
• A hook line. The first thing on screen.

What do you type into Vizard Agent?

State the finding the way the study did. Vizard Agent narrates what you give it, so the difference between "improves blood glucose regulation" and "is better than exercise" is a difference you have to make in the brief rather than one it can catch afterwards.

Prompt

Variants worth knowing:

• A hook badge in the upper third. Above the caption band, where it stays readable.
• A named source on screen. It changes how the claim reads.
• A calm narrator. Urgency in the voice undermines a scientific claim.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real health fact video. The last third of it is almost entirely about the two things that actually carry this format: the caption styling, and exactly where on the frame the hook badge is allowed to sit.

1. Checks the explainer template and the available tools.
2. Generates the voiceover and searches for background music.
3. Extracts word-level timestamps from the narration.
4. Searches and downloads high-resolution vertical stock footage for every scene, across two passes.
5. Tiles the clip thumbnails and reviews all the footage before choosing.
6. Generates styled animated captions and renders a test frame to check the placement.
7. Generates alternative caption styles and tests each on frame, comparing them at real timeline timestamps.
8. Measures the loudness of the narration and the music and builds the final captions with a vibrant highlight.
9. Renders the animated hook badge, composites it on a test frame, tries it in the upper third, verifies that position, then renders the full cut and reviews a frame grid across the whole timeline.

Step 9 is worth copying for any hook-led format. The badge is what earns the first two seconds, and it competes for space with the captions, so testing its position on a real frame rather than in a layout is the only way to know it works.

What does the result look like?

From the run this page is written from, probed on the delivered file: 1080x1920, H.264, 30fps, 59.5 seconds, AAC audio. Vertical, just under a minute, sourced footage cut to a narrated finding, with an animated hook badge in the upper third and highlighted animated captions below it.

Just under sixty seconds is the right size for a claim that needs context. Vizard Agent gives the finding room to be qualified, because a health fact compressed to fifteen seconds usually loses the conditions that make it true.

When does this not work well?

Health content is the one category where a well-made, convincing video does the most damage if the claim underneath it happens to be wrong. Vizard Agent narrates the finding exactly as you have stated it and does not evaluate, verify or qualify the science behind it.

• Vizard Agent does not verify the study. It narrates your wording; the evidence is yours to stand behind.
• A finding is not advice. What a study showed and what someone should do are different statements.
• Health claims are regulated. Many markets restrict what may be said and by whom.
• Precision gets lost in compression. The conditions on a finding are the first thing a short video drops.
• Sourced footage implies things. A clip of someone exercising can imply a claim the narration never made.

How do you fix a result that came back wrong?

Correct the wording first. Vizard Agent keeps the narration with its word timings, every footage candidate it downloaded, all the caption tests and both badge positions, so a rephrased claim is a re-record and a re-time rather than a rebuild of the video.

• "That overstates the finding." Reworded and re-recorded, and the captions follow.
• "The badge covers the subtitle." Repositioned in the upper third and verified, as here.
• "That clip implies the wrong thing." Swapped from the candidates already downloaded.

How does Vizard Agent compare to doing it yourself?

By hand this means writing the script, licensing footage, and setting captions in whichever style your editor defaults to. The hook card is usually an afterthought placed wherever there is room, which is why so many fact videos lose people before the claim has even been stated.

Common questions

Will Vizard Agent check the study? No. It narrates the finding as you word it, and the evidence is yours to stand behind.

How precisely should I state the claim? Exactly as the study did. Vizard Agent uses your wording, so any looseness in it survives into the video.

How long should a health fact video be? Around a minute, so the conditions on the finding survive.

Can Vizard Agent name the source on screen? Yes, and doing so changes how the claim reads to a viewer.

Why does the hook badge get so much attention? Because it is the first thing on screen and it competes with the captions for the same space. A badge placed by eye either overlaps the subtitle band or sits too high to read, and both cost you the opening seconds this format depends on.

Does Vizard Agent write the script? Yes, from your statement of the finding and the audience.

Can Vizard Agent make a shorter cut too? Yes, and be careful: the qualifications are what get cut first.

Does Vizard Agent check the finished video? Yes. It measures the final duration and loudness and reviews a frame grid across the whole timeline.