Brand TVC
A polished commercial assembled from consecutive scenes. The opening can be a person lip-synced to a line you recorded yourself.
Two TVC styles
| Style | Built around | Best for |
|---|---|---|
| Personal brand | One face. Opens with the presenter speaking, then moves to content scenes. | Shop owners, KOLs, anyone who appears on camera themselves. |
| UGC | The product, filmed as a real customer would. | Social content that needs to feel everyday. |
The UGC style also has its own page with a simpler flow if you just want a quick render without the full TVC screen.
Image inputs
Images accept PNG, JPG or WEBP, up to 10 MB each.
| Slot | Used for | Required |
|---|---|---|
| Character image | The face that appears in the commercial. | Yes, for the personal-brand style |
| Product image | The product being promoted. | Yes |
| Background image | The setting behind the video. | Yes, for the personal-brand style |
The product name has to be a real name
The Product name field is not just a label — the AI uses it to understand what is being advertised. "Dior Sauvage perfume" or "red lipstick" works; an internal code like "SP-0412" means nothing to it and produces the wrong scenes.
Writing the ad content
This describes shots, not dialogue
The content field is for describing the shots you want. The spoken words come from the voice settings. Write dialogue here and the AI only uses it to build imagery, never to speak it — this is the most common misunderstanding with this tool.
- Turn on Auto storyboard to have the system split your content into scenes. Turn it off and you describe each scene yourself.
- When writing them yourself, leaving a scene blank continues from the previous one — useful when you only care about a few key shots.
- Leave Video title blank and it takes the product name.
The opening voice
Upload a recording of your opening line. The person in the image is animated to speak that exact line, lip-synced, before the video moves on to the content scenes.
The opening eats into the total duration
The longer the recording, the shorter the content scenes, because both share the same total. Keep it to 3–5 seconds.
- This segment does not pass through the synthetic voice — the audience hears exactly the voice in your file. Record it in the voice you want them to hear.
- MP3 or WAV, clean, one speaker, no background music.
Audio
| Mode | Result |
|---|---|
| Voice-over | Narration generated from the promotional content, in the chosen voice's language. |
| Background music | Your uploaded track runs throughout, with no narration. |
| No audio | A silent video. |
Choosing background music requires an actual upload; otherwise the system prompts you before submitting.
Notes
- This tool has both Save to library and Publish now. Publish now skips the library and sends the video straight out.
- Product review, TVC and UGC share one rendering pipeline, so they share the same per-second billing.
- Short formats render as a single scene; only longer videos assemble several consecutive clips.