How to correct the names and places written on screen in a finished video

Last updated 2026-09-05 · 8 min read

Send Vizard Agent the list of corrections in plain sentences. It finds the existing label before rewriting anything, rebuilds only the name tags that are wrong, checks the corrected text at full resolution rather than in a thumbnail, and re-cuts the credit roll when somebody has been left off it.

What is the short version?

Only the people in the room know who somebody is and where they are from. A finished video will confidently caption both, and the corrections arrive as a short list of sentences that has nothing to do with editing and everything to do with facts.

Go to Vizard Agent with the delivered cut.
Say who each person actually is, in plain sentences.
Say what should be dropped and who is missing.

What do you need before you start?

The list, and nothing else. There is no need to describe where each tag appears or how it is styled — the existing labels get found in the video, and your job is only to say what is wrong with them.

Who each person is. Name, role, relationship to the event.
What to leave out. Surnames, titles, anything unwanted.
Where they are from, if that is on screen.
Who is missing. From the credits especially.
Anything that must not change. Usually the rest.

What do you type into Vizard Agent?

Write it as facts, numbered. This is not an edit brief and does not need to sound like one; a list of corrections is the clearest possible instruction because each line maps onto exactly one thing on screen.

Prompt

A few corrections. 1. The gentleman in the third interview is the owner of the venue, so his title is wrong. 2. Do not worry about anybody's surname — first names only. 3. She is from a different town than the caption says. And one last thing: I was conducting the interviews alongside my wife, so we should both be credited.

Variants worth knowing:

"First names only." A rule, applied everywhere.
"His title is wrong." Points at a tag without describing it.
"We should both be credited." Credits are the usual omission.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real vertical event vlog whose interview labels needed correcting after delivery. Notice that the existing label is looked up first, so a correction replaces a specific tag rather than adding a competing one.

Looks up the existing label already on screen.
Rebuilds the name tags that need to change.
Checks the new tags on frames from the render.
Fixes the specific title and hometown you identified.
Checks the corrected tags again on the new render.
Adds the missing interviewer credit to the ending.
Re-cuts the credit roll around the added line.
Rebuilds the ending with clearance so the credits fit.
Checks the credit text fits at full resolution.
Inspects the credit lines closely for clipped edges.
Uploads the corrected version and checks the tags in it.

Step one is the small discipline that prevents a mess. Rebuilding a tag from your correction alone risks changing its position, size or duration; finding the existing one first means the replacement inherits everything about it except the words.

Step seven is where a one-line correction turns into real work. Adding a name to a credit roll reflows every line beneath it, so the ending has to be re-cut around the new length rather than having the extra line squeezed into the space the old roll occupied.

Steps nine and ten are worth doing every time. Names are the text most likely to overflow — one longer surname, one extra line in the credits — and a thumbnail will show you the tag is present without showing you that its last letter is off the edge.

What does the result look like?

The same video, now with the facts right: every person titled correctly, hometowns as they actually are, surnames dropped wherever you asked for that, and a credit roll from Vizard Agent that names everyone who actually did the work.

Nothing else moves. The corrections are the only difference.

When does this not work well?

Corrections have to be findable in the first place. Text burnt into supplied footage rather than added by the edit cannot be rewritten by Vizard Agent, only covered over, and a name spoken aloud in the narration is a re-record rather than a text change.

Text inside your footage. Covering is the only option.
A name in the voiceover. That is a re-record.
Very long replacements. A longer name may not fit the tag.
Names in another script. The typeface has to support it.
Corrections without a subject. Say which person you mean.

How do you fix a result that came back wrong?

Say which tag is still wrong and what it should read instead. Vizard Agent keeps the label positions it looked up and the corrections it has already applied, so a second pass changes only those words rather than rebuilding the video around them.

"Still the old title." That tag rebuilt and re-checked.
"The name is cut off." Type refitted to the tag width.
"Wrong person tagged." The label moved to the right interview.
"A credit is missing." The roll re-cut with the line added.

How does Vizard Agent compare to doing it yourself?

By hand this is quick for one tag and error-prone across several, because each correction means finding the graphic, matching its style and re-exporting. The credit roll is the part that usually gets left, since adding a line reflows everything under it.

	By hand	Vizard Agent
Finding the tag	Scrubbing for it	Existing label looked up
Matching the style	Rebuilt from memory	Inherited from the original
Checking the fit	On a preview	At full resolution
The credit roll	Left as it is	Re-cut with clearance
The rest of the video	Re-exported	Untouched

Common questions

Do I need timestamps? No. Describe who you mean and Vizard Agent finds the tag.

Can it drop all the surnames? Yes. Give Vizard Agent the rule and it applies it throughout.

Will the style change? No. Vizard Agent reuses the existing tag's design.

What if the new name is longer? Vizard Agent refits the type and checks it at full size.

Can it add someone to the credits? Yes, and Vizard Agent re-cuts the roll to make room.

Does the whole video re-render? No. Vizard Agent rebuilds the affected sections and the ending only.

Can it correct captions too? Yes. Say which word and Vizard Agent fixes the subtitle.

What about a name said out loud? That needs the narration re-recorded; Vizard Agent will say so.

Can it handle accented characters? Yes, if the typeface supports them. Vizard Agent checks.

Will it list the tags for me? Yes. Ask Vizard Agent what each label currently says.

Can I send corrections in a batch? Yes, and Vizard Agent finds that the fastest way to work.

What if a name is burnt into the footage? Vizard Agent can cover it, but not rewrite it.

Does this work weeks later? Yes, as long as the project or the source is available.

Why look up the label first? Because a rebuilt tag that does not inherit the original's placement is a new graphic, and new graphics do not sit where the old one did.