Vizard Agent

How to make a LinkedIn video about a programme from your website

Last updated 2026-08-17 · 6 min read

Give Vizard Agent the page your programme sits on and say who it is for. Vizard Agent maps the site, reads the copy and testimonials, downloads the real photographs, samples the brand colours, auditions narrators until one reads it properly, and builds a square post video around your own material.

What is the short version?

A programme that exists as a page on your website is invisible on a professional feed. The material to promote it — the description, the photographs, the testimonials — is already written and already photographed; it just is not a video yet.

  1. Go to Vizard Agent and give it the page where the programme lives.
  2. Say what the programme is and who it is aimed at.
  3. Say the shape and roughly how long the post should run.

What do you need before you start?

The page and the audience. Vizard Agent maps your site and pulls what it needs, and what only you can supply is which audience this post is speaking to — teachers, parents or administrators produce genuinely different scripts from the same page.

What do you type into Vizard Agent?

Give the link and the context. Vizard Agent maps the site and finds the page even when it is nested, so a sentence naming the programme and its audience is worth more than a description of what the page already says.

Prompt

Make a [LinkedIn] video for my [literacy programme]. Here is my website — it is under [Resources]: [link]

Variants worth knowing:

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real programme post. The most instructive stretch is the casting: it did not take the first voice, and when a word came out mispronounced it tested spellings until the read was right.

  1. Looks at your website and maps its pages to find the programme.
  2. Reads the programme and resources pages, the copy and the testimonials.
  3. Collects the images from the site and downloads the real programme photographs, then looks at them.
  4. Samples the brand colours and generates the first voiceover, music and brand cards.
  5. Searches for voices, tests three narrator options, generates two more and compares them.
  6. Tests different spellings of a word to find which one the voice pronounces correctly.
  7. Records the final narration and takes word-level timings from it.
  8. Reviews the animated cards, downloads more photographs, checks the fonts and generates the captions.
  9. Assembles, reviews the frames, re-renders two cards, resizes the captions, balances the music and watches the finished video back.

Step 6 is the detail worth stealing. A generated voice mispronouncing a programme name is not fixable by asking it to try harder; changing how the word is spelled in the script is what changes the sound, and Vizard Agent tested until it found the spelling that worked.

What does the result look like?

From the run this page is written from, probed on the delivered file: 1080x1080, H.264, 30fps, 43.34 seconds, AAC audio. Square, forty-three seconds, narrated over your own programme photographs with animated brand cards, captions and the music balanced under the voice.

Vizard Agent watched the finished video back and then produced a critique of its own cut before delivering it, rather than treating the render as the end of the job.

Forty seconds is about six beats: what the programme is, who it is for, what it contains, the proof, the outcome and where to find it. That is the shape a professional feed post can carry.

Expect tens of minutes rather than minutes. This category was not separately measured; comparable work runs a median of 28 to 38 minutes end to end, and across all projects the median cost by tier is Flash 47, Pro 55, Max 242, Ultra 263 credits.

When does this not work well?

Everything in the video comes from your site and your own photographs. Vizard Agent maps the pages, gathers thoroughly and verifies what it can hear and see, and it inherits whatever is out of date, thin or unflattering in what your website currently publishes.

How do you fix a result that came back wrong?

Say which element. Vizard Agent keeps the site copy, the downloaded photographs, the narration, the word timings and each card separately, so re-rendering a card or re-recording one line does not repeat the site mapping or the casting.

How does Vizard Agent compare to doing it yourself?

By hand this is a screenshot session, a script written out from your own page, a voice you either record yourself or licence, animated cards built in a design tool, and a caption pass — all of it for one post, about one programme, on one platform.

By hand Vizard Agent
Finding the material Copy and paste from the site Site mapped, pages read
Photographs Hunt through the media library Downloaded and reviewed
The narrator Take what you get Five voices auditioned and compared
Pronunciation Live with it Spellings tested until correct

Common questions

Does it need me to send the copy? No. Vizard Agent maps the site and reads the page, including one nested under a section.

Will it use our real photographs? Yes. It downloads them from the site, and anything better you upload takes priority.

What if a name is mispronounced? Vizard Agent tests spellings to correct it, and you can flag names up front to save a round.

Can it match our brand colours? Yes. They are sampled from the site rather than guessed.

What shape should a professional post video be? Square works well in a feed. Vizard Agent can deliver vertical too, re-laid rather than cropped.

How long should it be? Forty seconds or so. Long enough for Vizard Agent to explain the programme properly, short enough to be watched at a desk.

Can it find a page nested under a section? Yes. Vizard Agent maps the site's pages rather than needing the exact deep link, so "it is under Resources" is enough to go on.

Can it use testimonials from the page? Yes. Vizard Agent reads them and can set them as cards; check anything attributed to a named person first.