Check captions against the source speech, then request the exact wording, timing, or placement change. Select the editable caption or title where that option is available. Review the new result for spelling and whether the text hides an important part of the picture.
The current wording and exact replacement spelling.
Text hides the subject
The affected scene and area that should stay visible.
Text moves too quickly
The phrase and timing change to review.
Hard-to-read style
A concrete colour, size, contrast, or position change.
What if the text is already in my footage?
Generated titles and captions can be editable elements in supported compositions. Text and objects already flattened into source footage are part of that video; selecting them does not automatically create separate layers.
Keep a clean source clip if you have one. Moving a separate caption is different from removing words burned into a video.
Try a precise correction
Change the editable caption “Acme Coffe” to “Acme Coffee”. Keep its timing and position the same. Leave other captions and surrounding shots unchanged.
How should I check the result?
Watch the revised phrase with sound, then check the downloaded file on a phone. Review names, numbers, line breaks, contrast, and product or face visibility.
Request its exact spelling, targeting the editable caption where supported, then review the new output.
Can every visible word be selected as text?
Generated titles and captions can be editable elements in supported compositions. Text and objects already flattened into source footage are part of that video; selecting them does not automatically create separate layers.
Ask an assistant
Take this page into your chat.
Open it in ChatGPT or Claude with the page already loaded, or read the plain-Markdown version.