How to make an explainer about a traditional art form

Last updated 2026-08-24 · 7 min read

What is the short version?

An explainer about a craft has to look like it respects the craft. That means real footage of the work rather than generic b-roll, and typography that a native reader recognises as correct rather than as software output.

1. Go to Vizard Agent and write the explanation in a short paragraph.
2. Give it a photograph of the presenter and the language.
3. Say the length and where the video will run.

What do you need before you start?

A paragraph and a photograph. Vizard Agent sources the footage of the craft, records the narration and builds the titles, so this format runs from very little. The language is the detail that shapes the most work, because it decides the voice, the typeface and how the text is laid out.

• The explanation. A short paragraph in your own words.
• A presenter photograph. Front-facing and clear.
• The language. For the narration, the captions and the titles.
• The length. Fifteen to twenty seconds for a single idea.
• The shape. Vertical for social.

What do you type into Vizard Agent?

Write the paragraph you would say out loud. Vizard Agent narrates from your text and builds every scene against its word timings, so a clear paragraph in the language you want is both the script and the structure at once.

Prompt

Variants worth knowing:

• A digital presenter from a photograph. Rather than a faceless narration.
• Traditional music. It carries the register better than generic score.
• Golden or ornamented titles. They suit heritage subjects and need real typefaces.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real heritage explainer. Nearly a third of the job is typography, because a script where letters join is either rendered correctly or is visibly wrong, with nothing in between.

1. Opens the attached photograph and reviews the available generation tools.
2. Checks the talking-photo settings and the narration options for the language.
3. Searches for a suitable male voice and generates the narration in formal register.
4. Measures the audio duration and extracts precise word timings to plan the scenes.
5. Generates the digital presenter and searches for footage of the craft — the patterns, the architecture, the calligraphy.
6. Measures the audio levels for a balanced mix.
7. Checks the installed typefaces, then downloads two proper ones for the script and confirms they are ready.
8. Generates the subtitle file and renders a test frame half a second in to check the letters join correctly.
9. Assembles the video, reviews a grid of frames, adjusts the title design and position, re-exports at a larger and clearer type size, and reviews the frames again.

Step 8 is the check that decides whether this video is publishable. A typeface that renders each character separately produces text that is technically present and unreadable as language, and the only way to know is to look at a rendered frame.

What does the result look like?

From the run this page is written from, probed on the delivered file: 1080x1920, H.264, 30fps, 18.0 seconds, AAC audio. Vertical, eighteen seconds, a presenter driven from a single photograph cut against real footage of the craft, narrated formally, with correctly rendered captions and titles.

Eighteen seconds is one idea explained properly. Vizard Agent kept it to a single definition rather than covering the history as well, because a short explainer that tries to be a documentary teaches nothing.

When does this not work well?

Cultural and religious subjects deserve a kind of care that no editing process can supply on its own, however well it is executed. Vizard Agent handles the language, the typography and the register properly, and the substance is still worth having reviewed by someone inside the tradition.

• Have the wording reviewed. Someone who knows the subject should read it before it goes out.
• Generated imagery of sacred subjects is sensitive. Sourced footage of real work is safer and better.
• Scripts with joined letters break easily. Vizard Agent tests them, and unusual fonts stay a risk.
• A photograph is not the person. A digital presenter should not be presented as filmed footage.
• Eighteen seconds is a definition. History, regional variation and technique each need their own video.

How do you fix a result that came back wrong?

Name the line or the title that is wrong. Vizard Agent keeps the narration timings, both downloaded typefaces, every test frame and all the sourced footage, so a rewritten sentence or a resized title is a targeted re-render rather than a rebuild of the video.

• "The wording is not quite right." Corrected and re-narrated from your text.
• "The letters are not joining." Another typeface, tested on a frame first.
• "The title is too small." Re-exported larger, as this run did.

How does Vizard Agent compare to doing it yourself?

By hand this means recording a narration, licensing footage, and then discovering that the caption tool renders your language as disconnected characters. Fixing that means hunting for a font that shapes correctly and hoping your editor supports it, which is where most people give up and post it with the text wrong.

Common questions

Do I need to film a presenter? No. Vizard Agent drives one from a single photograph.

Will the text render correctly in my language? Vizard Agent downloads proper typefaces and renders a test frame to check the letters join before building the video.

How long should an explainer like this be? Fifteen to twenty seconds for one idea. Longer subjects need a series.

Can Vizard Agent find footage of the craft itself? Yes, and it is worth insisting on. Real work beats generic b-roll for this subject.

Why does the typeface matter so much here? Because in scripts where letters connect, a font that renders them separately is not a styling problem but a legibility one. To a native reader the text looks broken, and everything else in the video inherits that impression.

Can Vizard Agent narrate formally? Yes. Name the register and Vizard Agent casts and directs the voice accordingly rather than defaulting to a conversational read.

Does Vizard Agent add music? Yes, and traditional instrumentation suits this register better than a generic score.

Should I have the wording checked? Yes. Vizard Agent narrates what you write, and cultural subjects deserve a reader from inside the tradition.

Can Vizard Agent make a series? Yes. One idea per video is the right unit, so a subject with several parts becomes several videos.

Does Vizard Agent check the finished video? Yes. It reviews frame grids twice and re-exported this one after enlarging the type.