How to show the article you are citing on screen

Last updated 2026-09-01 · 7 min read

What is the short version?

There is a difference between a video that says "according to the official figures" and a video that puts the page on screen while it says it. The second one is checkable, and on a platform full of confident voices, checkable is the whole argument.

1. Go to Vizard Agent and paste your script.
2. Give it the link to the article or report you are citing.
3. Say you want the source shown on screen, not just mentioned.

What do you need before you start?

Very little beyond the argument itself. Vizard Agent supplies the narration, the footage bed and the graphics, so the two things only you can bring are the point of view you are actually making and a link to whatever it is you are leaning on for authority.

• Your script. The argument in your own words.
• The link. The article, report or data page.
• The claim it supports. Which line it backs up.
• The platform. It sets the frame and the safe area.
• The tone. Explainer, commentary or straight news read.

What do you type into Vizard Agent?

Say the word "showcase" or "show it on screen" explicitly. Left to itself, Vizard Agent will happily read a source out in the narration and move on, because a citation is a fact about the script rather than a visual instruction — you have to ask for it as a picture.

Prompt

Variants worth knowing:

• "Showcase the article." Turns a citation into a shot.
• "Cards for the numbers." Each figure gets its own beat.
• "End on the question." The closing line as its own card.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real property-market commentary built from a written script and one linked article. Step four is the one people do not expect: it looks at its own screenshot before deciding the page is usable.

1. Checks the explainer template guidelines and the available tools.
2. Fetches the linked article and reads what is actually on the page.
3. Screenshots the article as it renders in a browser.
4. Looks at the screenshot to confirm the headline is legible.
5. Checks the motion-graphics options and the caption presets.
6. Auditions voices for a news-explainer read.
7. Generates the narration and searches for a music bed.
8. Transcribes its own narration to get word-level timings.
9. Re-generates the narration where a phrase did not read cleanly.
10. Searches stock footage scene by scene against the script.
11. Builds the article showcase card as an animation.
12. Extracts a frame from the card and inspects it.
13. Builds the remaining cards — the figures, the link, the closing question.
14. Tiles every card into one preview grid and reviews them together.
15. Measures the loudness of narration against music.
16. Generates animated subtitles and moves them into the safe zone.
17. Renders a ten-second slice to check the assembly before the full run.

Steps three and four are the heart of it. A screenshot of a page is not automatically a usable shot: paywalls, cookie banners and lazy-loading images all produce a picture of a page that says nothing. Vizard Agent looks at the image it captured, the same way you would, before it builds anything on top of it.

What does the result look like?

The video this page is written from opens on the argument, and when the narration reaches the source, the article itself animates onto the screen — the real headline, the real page, captured from the live URL rather than mocked up in a design tool. The figures then land as their own cards as they are spoken.

The effect is quiet and it does a lot of work. The viewer is not asked to trust a voice; they are shown the thing the voice is reading from, which is the difference between commentary and an opinion.

When does this not work well?

A screenshot is a photograph of whatever the browser was showing at that moment, and a great many pages are not showing their content at that moment. These are the cases where the card is better replaced with something else, and where Vizard Agent will tell you rather than shipping a picture of a cookie banner.

• Paywalled articles. The page shows a wall, not the piece.
• Login-gated reports. Nothing loads without a session.
• Long PDFs. A page of dense text is unreadable at video size.
• Pages that block screenshots. Some do; supply an image instead.
• Sources you may not reproduce. Licensing is yours to check.

How do you fix a result that came back wrong?

Say what is wrong with the card and Vizard Agent rebuilds only that. The narration, the word timings, the loudness measurements and the rest of the cards all survive, because the article card is one animated layer in the timeline rather than something baked into every frame of the render.

• "The headline is cut off." Re-captured at a different width.
• "It sits on my face." Repositioned clear of the subject.
• "Wrong part of the page." Cropped to the section that matters.
• "It is on screen too long." Retimed to the line it supports.

How does Vizard Agent compare to doing it yourself?

By hand this is a browser, a screenshot key, a crop, a keyframed zoom and a guess about how long to leave it up. None of that is hard, and all of it is fiddly enough that most people skip it and let the narration carry the claim on its own.

Common questions

Does it need the article to be public? Yes, for a live capture. Vizard Agent tells you when a page will not load.

Can I supply my own screenshot instead? Yes. Upload the image and Vizard Agent builds the card from that.

Will it read the article for me? It can. Vizard Agent pulls the page content as well as capturing the picture.

Can it show more than one source? Yes. Vizard Agent builds a card per source and spaces them across the script.

What about a PDF report? Give it the page or a screenshot of the relevant table. Full PDFs do not read at video size.

Does it check the numbers in my script? Only if you ask. Vizard Agent can look the figures up against the source you linked.

Can the card be a lower third instead? Yes. Vizard Agent will keep it small and off to one side if you say so.

How long should it stay up? Long enough to read the headline. Vizard Agent times it to the line it supports.

Will it cover my captions? No. Vizard Agent places the subtitles inside the safe zone and works around them.

Can it highlight one sentence in the article? Yes. Say which line and Vizard Agent crops and emphasises that part of the capture.

Does it work for a horizontal video? Yes. The frame changes the crop; Vizard Agent captures to fit whatever ratio you asked for.

Can I reuse the card in the next episode? Yes. Vizard Agent keeps the built card and can re-time it to a new script.

What if the site redesigns after I publish? Your capture is already rendered into the video. Vizard Agent works from the page as it was on the day.

Does it credit the publication? It shows the page, which carries the masthead. Ask for a text credit as well if you want one on screen.