How to make an introduction video for a subject you only named
Give Vizard Agent a name and a length. Vizard Agent researches what the subject actually is, sources footage for it, generates titles, assembles a ten-second introduction — and when the sourced shots do not show the real thing convincingly, it generates the missing footage rather than settling for an approximation.
What is the short version?
Sometimes the whole brief is a name and a duration. That is enough, provided the first step is finding out what the name refers to — because a video built on the wrong reading of a subject is wrong in every frame.
- Go to Vizard Agent and name the subject.
- Say the length, the shape and where it is going.
- Say the feel if you have one; otherwise Vizard Agent picks from the subject.
What do you need before you start?
A name and, ideally, a sentence of context. Vizard Agent will research an unfamiliar subject and work out whether it is a dish, a person, a mascot or a place — and a one-line clarification from you removes the only real risk in a brief this short.
- The subject. A dish, a brand, a character, a place, a mascot.
- A word of context. "It is a restaurant" saves a research step and a possible misreading.
- The length. Ten seconds is the standard for an introduction card.
- The shape. Wide for a channel opener, vertical for a feed.
- Any footage you have. Vizard Agent uses yours before sourcing or generating.
What do you type into Vizard Agent?
A single line genuinely works. Vizard Agent researches the subject, decides what needs to be shown, sources or generates it and writes the titles, so the brief only needs to grow if the name is ambiguous or you want a particular angle.
Prompt
Variants worth knowing:
- Add the angle. "Focus on how it is served" narrows what gets shown.
- Your own footage. Upload it and Vizard Agent builds around your material.
- A set of them. Several subjects in the same style makes a series rather than one card.
What does Vizard Agent actually do?
Here is the order Vizard Agent worked in on a real introduction video. The beginning and the end both matter: it finds out what the subject is before sourcing anything, and it refuses to accept a stock clip that does not show the real thing.
- Searches for information about the subject to understand what it is.
- Refines the search to establish whether it is a person, a place or something else.
- Searches for high-quality footage of the subject and its context.
- Reads the stock tool's options and runs more specific searches.
- Downloads the selected clips, including one to match the others' shape.
- Generates the title graphics, then regenerates them with transparency and in the right aspect ratio.
- Assembles the clips, overlays and music into the ten-second cut.
- Uploads it and builds a tiled preview, then samples frames across the whole clip and checks the timing.
- Searches again for a more authentic version of one shot, generates it when the library has nothing true enough, re-assembles, and reorders the sequence.
Step 9 is the judgement worth noticing. Vizard Agent looked at its own finished video, decided one clip did not genuinely show the subject, and generated a replacement rather than shipping something that was close enough to pass.
What does the result look like?
From the run this page is written from, probed on the delivered file: 1920x1080, H.264, 30fps, 12.5 seconds, AAC audio. Wide and twelve seconds, mixing sourced footage with a generated shot, animated titles composited over the picture with transparency, and music underneath the whole thing.
Vizard Agent verified the result by sampling frames across the whole duration rather than checking the opening, which is how the weak shot was caught in the first place.
How long it takes was not measured separately here. Comparable work runs a median of 28 to 38 minutes end to end, and the upper quartile is several times the median. Across all projects the median cost by tier is Flash 47, Pro 55, Max 242, Ultra 263 credits.
When does this not work well?
Ten seconds about a named subject depends entirely on getting the subject right and finding footage that genuinely shows it. Vizard Agent researches and verifies, and there are subjects where the honest answer is that no library has the right material.
- An ambiguous name is a real risk. Add a word of context if the name could mean two things.
- Stock rarely shows the specific thing. It shows the category. For anything particular, generation or your own footage is closer.
- Research can be wrong. Vizard Agent grounds the video in what it finds; check anything factual before publishing.
- A real business needs real footage. For your own restaurant or shop, film it. Nothing sourced is your place.
- Ten seconds is three shots. It introduces; it does not explain.
- Titles do the talking. With no narration, whatever the cards say is the entire message, so the wording matters more than the footage does.
How do you fix a result that came back wrong?
Say what is off. Vizard Agent keeps the researched context, every downloaded and generated clip and the title graphics, so replacing a shot or reordering the sequence is an assembly change rather than a fresh round of searching.
- "That is not what the subject is." Re-researched and rebuilt with the correct reading.
- "That clip is not authentic." Replaced by a generated shot of the real thing.
- "Put the close-up first." A reorder; the titles follow.
How does Vizard Agent compare to doing it yourself?
By hand this is a search engine, a stock site, a titles template and an editor — plus the moment where you accept a clip that is nearly right because finding a better one would take another hour.
| By hand | Vizard Agent | |
|---|---|---|
| Understanding the subject | Assume you know | Researched first |
| Sourcing shots | Take what the library has | Searched, then generated if lacking |
| Titles | A template with text typed in | Generated with transparency, to shape |
| Checking | Watch the opening | Frames sampled across the whole clip |
Common questions
Do I need to explain what the subject is? Not usually. Vizard Agent researches it, and a word of context helps when the name is ambiguous.
What if there is no good stock footage? Vizard Agent generates the missing shot rather than using a clip that does not show the real thing.
Can it use my own footage? Yes, and it prefers it. Upload what you have and Vizard Agent builds the introduction around your material, sourcing only what is missing.
How specific can the subject be? Very. Vizard Agent researches it first, and a word of context removes any ambiguity in the name.
How long should an introduction be? Ten seconds. It exists to introduce, and anything longer needs a reason to keep watching.
Can I get a vertical version? Yes. Vizard Agent regenerates the titles for the new shape rather than cropping the wide ones, so the text stays inside the frame.
Does it add music? Yes. Vizard Agent searches a licensable library and picks a track that suits the subject rather than defaulting to one.
Can it make a series of these? Yes. Give several subjects and Vizard Agent keeps the titles, the pacing and the music style consistent across them.
Can it add narration? Yes, though ten seconds rarely needs it. Titles usually carry an introduction better than a voice does.