> ## Documentation Index
> Fetch the complete documentation index at: https://docs.superscale.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Creating videos

> Generate, edit, and combine AI video, B-roll, scripts, voiceover, and captions into finished ads ready to export.

Superscale produces finished video ads from a brief: scripted, produced, voiced, captioned, and ready to export. The strongest production workflows are usually hybrid: use AI for hooks, storyboards, creators, or generated scenes, then combine them with real product footage, screenshots, B-roll, captions, and controlled edits.

<Frame>
  <img src="https://mintcdn.com/superscale/N3uCUOnb7HvtowwW/images/Capture-2026-06-11-184239.png?fit=max&auto=format&n=N3uCUOnb7HvtowwW&q=85&s=9740c327ba91c2e3acbbdc06c55276df" alt="Capture 2026 06 11 184239" width="1449" height="1206" data-path="images/Capture-2026-06-11-184239.png" />
</Frame>

## Choose the right video workflow

<CardGroup cols={2}>
  <Card title="Seedance chat video">
    Best for fast 9:16 paid-social concepts generated directly from a brief, storyboard, and references.
  </Card>

  <Card title="UGC testimonial">
    A synthetic creator talks to camera in English. Strong for direct-response testimonial-style ads.
  </Card>

  <Card title="Voiceover with B-roll">
    Best when product footage, app screens, packaging, or screen recordings should be the visual proof.
  </Card>

  <Card title="Video templates">
    Use the reusable template/avatar workflow when you need a saved actor, actor selection, or more controlled assembly.
  </Card>
</CardGroup>

## How a video comes together

<Steps>
  <Step title="Describe the ad" icon="message-square">
    Include the audience, angle, format, language, product, and any references. If you do not know the angle yet, ask the agent to research options before it writes the storyboard.
  </Step>

  <Step title="The agent writes the storyboard" icon="file-text">
    It turns your brief into a [scene-by-scene script](/features/scripts), grounded in your context, references, and research signals.
  </Step>

  <Step title="Generate scenes and voice" icon="video">
    One storyboard becomes a multi-scene 9:16 video, with native English audio or a post-mixed [voiceover](/features/voiceover-lipsync) for other languages.
  </Step>

  <Step title="Export or publish" icon="crop">
    Export the finished asset. If Superscale shows resize or publish controls for that asset and connected destination, use them from the export view.
  </Step>
</Steps>

## Seedance formats

Seedance chat video is currently **9:16 vertical**. The format picker offers the formats that fit your product type and language.

| Format                       | Best for                                                                                               | Language notes                                        |
| ---------------------------- | ------------------------------------------------------------------------------------------------------ | ----------------------------------------------------- |
| **UGC Testimonial**          | Selfie-style creator talking straight to camera.                                                       | English only, native audio.                           |
| **UGC Lifestyle**            | Person on screen using, unboxing, scrolling, cooking, or applying while narration carries the message. | All languages. Non-English uses post-mixed voiceover. |
| **B-Roll (Faceless)**        | Product close-ups, hands, overhead shots, macro details, and app/product proof without a face.         | All languages. Physical-product focused.              |
| **POV (First-Person)**       | Hands and first-person view, useful for TikTok-native product moments.                                 | All languages.                                        |
| **Before / After**           | Same-frame transformation or contrast structure.                                                       | All languages.                                        |
| **Animated Brand Character** | Product, package, logo, or app icon comes to life as a character.                                      | English only.                                         |
| **Cinematic Brand Film**     | Premium product film, calm narration, no creator on camera.                                            | All languages.                                        |
| **Studio Tabletop**          | Product theatre on a studio backdrop.                                                                  | All languages, physical-product focused.              |
| **Slice-of-Life Cinematic**  | Polished lifestyle scene with a person on screen who does not speak.                                   | All languages.                                        |
| **Stop-Motion**              | Objects arrange around a physical product on a simple backdrop.                                        | All languages, physical-product focused.              |
| **Hands-Only Demo**          | Step-by-step physical product tutorial.                                                                | All languages, physical-product focused.              |
| **Mock Infomercial**         | Exaggerated problem-solution parody for irreverent brands.                                             | All languages.                                        |
| **Long Form Story**          | Longer narrative ad where the agent resolves the best matching style first.                            | Beta, 20-45 seconds.                                  |

<Note>
  Chat video generation does not currently create horizontal or 16:9 Seedance videos. If you need a different final placement, start with a vertical ad and use the export or recreation controls shown for that asset.
</Note>

## Before spending video credits

Video is expensive compared with statics, so reduce uncertainty first.

* Write or approve the script before rendering.
* Attach exact product visuals, app screenshots, or B-roll if they matter.
* Add pronunciation guidance for brand names and acronyms.
* Use a short test or smaller batch before scaling a format.
* Keep the output goal clear: hook test, product demo, testimonial, localization, or winner iteration.
* Decide whether you need Seedance chat video or a reusable video template/avatar workflow before briefing the asset.

## Common video questions

<AccordionGroup>
  <Accordion title="Should I generate everything from scratch?" icon="clapperboard">
    Not always. Use real B-roll, screenshots, product footage, and testimonials when those details need to be accurate. Use AI for the parts you cannot shoot quickly.
  </Accordion>

  <Accordion title="Can I edit one scene without remaking the whole video?" icon="square-pen">
    Think scene-by-scene. If one scene is weak, use the scene regeneration or replacement controls shown for that asset instead of starting from zero.
  </Accordion>

  <Accordion title="How long should the video be?" icon="timer">
    Most Seedance formats default to short-form, usually 15 seconds. Supported chat durations include 4, 5, 8, 10, 12, 15 seconds, and longer 20-45 second story formats when those formats are enabled in your video workflow.
  </Accordion>

  <Accordion title="Can I make a non-English talking-head testimonial?" icon="languages">
    Not as a native talk-to-camera Seedance format. Use UGC Lifestyle or B-Roll with post-mixed voiceover for non-English ads.
  </Accordion>

  <Accordion title="What if the video looks too AI?" icon="eye">
    Add real assets, use more natural scripts, reduce over-polished delivery, choose a better voice, and keep scenes tighter. Hybrid workflows usually look more believable.
  </Accordion>
</AccordionGroup>

<Card title="Set up video context" horizontal href="/features/video-context">
  Give the agent the footage, references, voice rules, and product details it needs.
</Card>