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

# Generate an AI Short-Form Video from Any Topic Idea

> Go from a topic idea to a fully composed short-form video — AI script, visuals, voiceover, and captions — without leaving your browser. Takes just minutes.

This guide is a complete walkthrough of the **Shorts Generator** in Origami AI. You'll enter a topic, configure your short, review an AI-generated scene-by-scene script, produce visuals and voiceovers in parallel, pick a caption style, and render a finished video — all from one workflow.

<Steps>
  <Step title="Navigate to Shorts">
    Click **Shorts** in the Origami AI navigation menu. This opens the Shorts composer, separate from the main PDF-to-video workflow.
  </Step>

  <Step title="Set up your short">
    Fill in the configuration panel before generating a script:

    | Option              | Description                                                              |
    | ------------------- | ------------------------------------------------------------------------ |
    | **Topic / Concept** | A sentence or phrase describing what your short is about                 |
    | **Duration**        | 15–90 seconds — shorter videos render faster and perform well on social  |
    | **Tone**            | Educational, entertaining, inspirational, or similar                     |
    | **Aspect Ratio**    | 9:16 for TikTok/Reels/Shorts, 16:9 for YouTube/landscape, 1:1 for square |
    | **Visual Style**    | Controls the aesthetic of AI-generated images and clips                  |

    Take a moment to set these carefully — they shape every downstream step.
  </Step>

  <Step title="Review the script">
    Origami AI generates a **scene-by-scene script** with two parts for each scene:

    * **Voiceover text** — what the narrator says
    * **Visual prompt** — the description sent to the image/video generator

    Read through every scene before proceeding. Click any scene card to edit the voiceover text or visual prompt. Getting this right now saves you from regenerating individual scenes later.
  </Step>

  <Step title="Generate assets">
    Click **Generate** to kick off parallel asset production for all scenes:

    * **Visuals** — images or short video clips are produced via [Pollinations](https://pollinations.ai) using your visual prompts
    * **Voiceover** — Kokoro TTS synthesizes speech audio for each scene's voiceover text locally in your browser

    Progress cards show the status of each scene. You can watch scenes populate in the storyboard as they complete — no need to wait for all of them.

    <Note>
      Visual generation requires an active internet connection. Pollinations is an external service, so image and video clip production happens in the cloud, not on your device.
    </Note>
  </Step>

  <Step title="Choose captions">
    Select a caption style that fits your brand and platform:

    | Style                 | Character                                   |
    | --------------------- | ------------------------------------------- |
    | **Bold Pop**          | Large, punchy words that animate on screen  |
    | **Karaoke Fill**      | Words highlight in sync with the voiceover  |
    | **Clean Lower Third** | Subtitle-style bar at the bottom            |
    | **Highlighter**       | Text with a colored background highlight    |
    | **Neon Glow**         | Glowing neon text effect                    |
    | **Classic Cinema**    | White text on a semi-transparent black band |

    After choosing a style, adjust the **font size** and **vertical position** to suit your aspect ratio.
  </Step>

  <Step title="Add music (optional)">
    Open the **Music** panel to select a background track. Set the volume so it sits comfortably under the voiceover — a level around 20–30% works well for most tones. Leave music off for a cleaner, voice-focused short.
  </Step>

  <Step title="Export">
    Click **Render** to compose the final video on your device. Origami AI stitches together your visuals, voiceover audio, captions, and music using client-side encoding — no files leave your machine during rendering. The finished MP4 downloads automatically when complete.
  </Step>
</Steps>

## Connecting Pollinations

Visual generation in Shorts is powered by [Pollinations](https://pollinations.ai). By default, Origami AI routes requests through the free tier, which works without an account but has lower rate limits and may be slower.

To get higher limits and access to additional models:

1. Open **⚙️ Settings → API**.
2. Scroll to **Pollinations Account** and click **Connect with Pollinations**.
3. Sign in and grant token access. Your credentials are saved locally in your browser — never on Origami AI's servers.
4. If your token expires, click **Reconnect** to refresh it.

If you don't have a Pollinations account or prefer not to connect one, select the **Free (slow model)** option in the model dropdown within the Shorts composer. It's slower but always available.

**Available visual models include:**

<Tabs>
  <Tab title="Images">
    `flux`, `zimage`, `nanobanana`, `krea`, and more — check the model dropdown for the current list.
  </Tab>

  <Tab title="Video clips">
    `wan-fast`, `seedance-2.0-fast`, and similar — short animated clips that add motion to your scenes.
  </Tab>
</Tabs>

## Regenerating Scenes

You don't have to rebuild the entire project if one scene doesn't turn out right. Click the **regenerate icon** on any individual scene card in the storyboard to re-generate just that scene's visual, voiceover, or both — all other scenes stay untouched. This is the fastest way to iterate on problem scenes without losing work you're happy with.