AI presenter
One presenter photo and a topic become a studio presentation, 10 seconds per scene, with the presenter speaking the lines in sync.
Steps
- 1
Open AI presenter
Go to Create video, group Avatar & motion, and pick AI presenter.
- 2
Pick the presenter photo and setting
Under Presenter & setting, choose exactly one presenter photo from the library or upload a new one, then pick the studio setting and turn holograms on or off.
- 3
Choose how it is shot
Under Camera & transitions, choose the camera movement, aspect ratio and transition style.
- 4
Write the topic and split scenes
Under Content & scenes, pick a content style, enter the topic, set the scene count and click AI splits … scenes & writes dialogue. Edit any scene you want.
- 5
Check the cost and create
Read the credit breakdown and click Create and save to library. The finished video lands in the video library, where you publish it.
Options
| Option | Choices |
|---|---|
| Content style | Explainer, Product intro, Behind the scenes, Case study, Myth vs fact, Brand story, Tips / top list, Q&A, Industry news, Comparison |
| Setting | Production studio, Glass office, Showroom, Lab |
| HUD holograms | On or off; blue, gold, green or violet |
| Camera movement | Slow dolly in, Side dolly, Static, Orbit |
| Aspect ratio | Landscape 16:9 or portrait 9:16 |
| Transitions | Auto, Dissolve, Slide, Zoom, Flash, Blur |
| Dialogue language | Vietnamese or English |
Holograms are transparent interface panels floating beside the presenter, drawn by the model inside the shot. They are decoration and carry no text — do not rely on them to show figures.
The panel on the right is a layout preview of your current choices, not the actual video. Click scenes on the timeline to step through it.
Dialogue per scene
- The topic takes up to 300 characters. Scene count runs from 1 to 8, default 3.
- The first scene opens, the last one wraps up with a call to action, and the ones in between carry the main content. Each scene takes a shot size: wide, medium or close-up.
- Each scene takes up to 200 characters of dialogue, but 10 seconds only fits about 150. The counter under the field shows where you are.
- Leave a scene's dialogue empty and the model writes it at render time from the topic — so the topic cannot be empty while any scene is.
The voice comes from the video model itself
This tool does not use the voice library. The model renders picture and sound together, so the presenter speaks already in sync. For one specific voice, use the talking AI avatar with your own audio file.
Duration and cost
Every scene is exactly 10 seconds, so the video runs scene count × 10 seconds: 3 scenes is 30 seconds, 8 scenes is 80. The form shows the "scenes × 10 seconds = total" line as you go.
The tool is billed per second on that total and runs on Server 1 only, so there is no server picker. Credits are held when you click create, charged when the video finishes and returned to your wallet if the render fails. See the rate in how credits work.
What makes a good presenter photo
- Clear face, looking straight at the camera, upper body visible.
- Wearing the outfit you want on screen — every scene keeps this face and outfit.
- One person in frame, even lighting, sharp image.
Notes
- Scenes are rendered separately and joined with the transition you picked. Expect a few minutes.
- The web form renders and saves to the library; it does not publish. Publish from the video library.
- AI assistants can use this tool over MCP: write_ai_presenter_script writes the dialogue, create_ai_presenter_video renders the video.