How to stop a name being spelled wrong in every clip you publish
Give Vizard Agent the exact spellings before it cuts anything. Speech recognition does not know how a surname is spelled, so it produces a plausible version. In a batch of clips, that version appears on every name label and every caption at once. Vizard Agent also checks which speaker sits on which side of the frame before any label goes on.
What do you type into Vizard Agent?
Put the spellings in the first message as a list, show title included and styled exactly. It takes one line and it is the difference between six correct clips and six corrections, and it gives Vizard Agent something to check the captions against as it goes. Asking to see one finished clip first is the batch check.
Prompt
What does Vizard Agent actually do?
Cutting one episode into six clips, each carrying two names and a show title on screen, turns one wrong spelling into six public mistakes. Putting the right names on the wrong sides is worse still, because it looks deliberate. That is why all the identification work happens before any clip is built. In the session behind this article it:
- Reviewed the episode for strong ten-to-thirty-second clips using the exact names supplied, after checking the conversation, camera layout and introductions.
- Read the transcript against the actual frames so each clip credited the right speaker, and confirmed which speaker sat on which side.
- Separated complete advice from answers that needed too much setup, locked six cut points and checked face positions at full resolution so the portrait layout cleared both faces.
- Built the shared titles and name labels, measured the clips' sound, added word-timed captions and rendered one layout check before exporting the batch.
- When the first export failed, read the render error, corrected an audio-format mismatch and checked the caption size.
- Reviewed the portrait crop, the exact spellings and the caption placement on that first export.
When does this not work well?
Some episodes can't be labelled reliably. If the speakers are never both in shot, swap seats partway through, or a guest is never introduced by name, the attribution has to come from you rather than from anything Vizard Agent can read. Overlapping speech turns attribution into guesswork, and several people with the same first name need telling apart.
Common questions
Why does speech recognition get names wrong? It hears sounds, not spellings, which is why Vizard Agent asks for them.
What if my show title has odd punctuation? Give Vizard Agent the exact form and it uses that everywhere.
Can I check one before the rest? Yes. Ask Vizard Agent for one finished clip first. It is worth doing every time.