How to keep your title text out of the generated image
Tell Vizard Agent that the titles are typography, not part of the picture. It builds each card as real type laid over the image rather than asking the generator to draw the words, so the spelling is yours. Then it measures what each finished card actually runs for, because a card asset routinely comes back longer than the duration it was asked for.
What do you type into Vizard Agent?
Tell Vizard Agent where the text actually lives. "Make me a title card" is ambiguous between a designed card and a generated picture of a card, and only one of those will spell your title correctly on every render. Give the exact wording, including punctuation.
Prompt
What does Vizard Agent actually do?
Typography laid over the picture stays under your control: the spelling, the face, the size, the position. None of that survives being described to an image generator. The second idea is measuring each card after it exists, because a timeline built by adding up the durations you asked for won't match the ones you got. This was a documentary explainer with an opening card, two chapter cards and a closing card. In the session behind this article it:
- Checked the generation options and their prices, verified the facts against institutional sources and wrote the narration first, so the edit was paced to the real spoken duration.
- Generated and measured the narration in one consistent voice, then the visual set with an instrumental bed, and reviewed the frames for consistency and artefacts.
- Created the opening card, each chapter card and the closing source card with exact typography, not embedded in the imagery.
- Built the picture timeline from the stills and the cards, and found the assembly was longer than intended because the cards carried their own durations.
- Measured each card's real duration rather than the one requested, trimmed the cards to their intended holds and brought the timeline back to length.
When does this not work well?
Text that has to sit inside the world of the shot is a different problem. A sign on a building, a label on a packet or writing on a screen is part of the image, and pulling it into a type layer means compositing it back into perspective. Handwriting and text that must look weathered are poor fits for flat type too.
Common questions
Why can a generator not spell? It draws letter shapes rather than setting text. Vizard Agent sets text.
Why did my timeline come out long? Card assets carry their own durations. Vizard Agent measures them.
Will regenerating the image break the title? No. Vizard Agent keeps the title on a separate layer.