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

# Create a Video

> Authorize JoggAI, describe a Smart Video project, and review the generated result.

## Start with a complete request

For the best first result, include:

* Topic or source material
* Target duration
* Output language
* Landscape or vertical format
* Talking-avatar placement
* Any required facts, assets, or brand guidance

```text theme={null}
Create a 3-minute landscape video explaining solid-state batteries in English.
Use a talking avatar only in the opening and closing scenes. Use B-roll and
motion graphics for the body, include subtitles, and keep the tone technical.
```

The requested duration is a planning target. Smart Video measures the completed narration and uses its real timing for scenes, subtitles, motion graphics, and the final render.

## Connect media services

Before planning a new video, Smart Video starts its local service and returns a ready Settings page. Choose one media source explicitly.

<Tabs>
  <Tab title="JoggAI OAuth">
    Select **Connect JoggAI** and complete authorization in the browser. JoggAI provides the voices and avatars visible to your account and Space.

    Smart Video stores authorization in macOS Keychain or Windows protected storage. Access and refresh tokens are not written to project files or shown to Codex.
  </Tab>

  <Tab title="Local Media">
    Select **Local Media** before generation starts to use supported on-device TTS and ASR. Talking-avatar generation requires JoggAI authorization.

    Local Media is a deliberate alternative, not an automatic fallback. A project that has already submitted JoggAI generation work will stay on JoggAI so paid work is not duplicated.
  </Tab>
</Tabs>

<Note>
  Do not open or bookmark a guessed `localhost` or `127.0.0.1` port. The port changes between sessions. Always use the Settings or editor URL returned after Smart Video reports that the service is ready.
</Note>

## Review the production flow

Smart Video turns the request into a structured brief, narration, shot strategies, visual scenes, and a local timeline. Depending on the plan, it can combine:

* Voiceover and timed subtitles
* Talking-avatar opening, closing, or full-video scenes
* Muted-by-default B-roll that can be unmuted in the editor
* HTML motion graphics and ECharts visualizations
* Background music and transitions

Motion-graphic scenes are generated and visually checked before media generation continues. If a scene needs correction, Codex repairs that scene without replacing already approved work.

## Open the project

When the project is ready, Codex returns an editor URL. The editor opens locally and defaults to English; Chinese is available from the interface language setting.

An incomplete project can still open for inspection and repair. Final MP4 export remains strict and waits until all required render assets are available.