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

# Generate from an Image

> Turn a still image into an animated video clip by describing the motion you want.

Image-to-video generation takes a still image as the starting frame and animates it based on your prompt. The AI preserves the visual content of your image and adds the motion you describe.

## How to generate a video from an image

<Steps>
  <Step title="Select Video as your artifact type">
    On the home dashboard, click the artifact picker and select **Generate Videos**.
  </Step>

  <Step title="Configure your settings">
    Set the aspect ratio and audio toggle. The same options apply as text-to-video generation.

    <Tip>
      Muted videos use half the credits. Turn audio off if you do not need spoken dialogue or ambient sound.
    </Tip>
  </Step>

  <Step title="Attach your starting image">
    Click the **+** button in the input bar and upload the image you want to animate. This image becomes the first frame of the video.

    <Frame caption="An image attached to the input bar, ready to be animated.">
      <img src="https://pub-51091dcf1e9d4b04bb2e74f489c4f346.r2.dev/5801e4db1861fd7cecea0b8b8c18ea50f4753c3a6439be597c57ac037df9d3d2.png" alt="Input bar with an uploaded image thumbnail visible, ready for a video generation prompt" />
    </Frame>
  </Step>

  <Step title="Describe the motion">
    Type a prompt describing what should happen in the video. Focus on motion, not on what the image looks like (the AI already has the image).

    <Prompt description="Animate a product photo for an Instagram Reel">
      I uploaded a photo of our coffee bag on a marble countertop. Slowly zoom in on the bag over 5 seconds. Add a subtle steam effect rising from behind it as if there is a fresh cup just out of frame. Warm lighting shifts slightly from left to right. 9:16 for Instagram Reels. Generate muted.
    </Prompt>

    <Prompt description="Bring a storefront photo to life for a website hero">
      I uploaded a photo of our gym's exterior at dusk. Animate the scene: turn on the interior lights gradually, add a few people walking past on the sidewalk, and add a subtle flicker to the neon sign. 8 seconds, 16:9. Generate with audio, city street ambiance.
    </Prompt>

    <Prompt description="Animate a food flat lay for a social media ad">
      I uploaded a top-down photo of our grain bowl. Animate the tahini drizzle slowly pouring across the bowl from top to bottom over 6 seconds. The rest of the food stays still. Warm natural lighting. 1:1 square. Generate muted.
    </Prompt>
  </Step>

  <Step title="Wait for generation">
    The agent generates the video using your image as the starting frame. A placeholder appears on the canvas while it processes.
  </Step>

  <Step title="Review and refine">
    The video appears on the canvas. Click to play. If the motion is not right, describe what to change or generate again with an updated prompt.
  </Step>
</Steps>

## A→B transitions

You can provide both a starting image and an ending image. The AI generates a smooth transition between the two states.

<Steps>
  <Step title="Attach both images">
    Upload two images: one for the start and one for the end. Mention in your prompt which is the start and which is the end.
  </Step>

  <Step title="Describe the transition">
    Explain what changes between the two frames.

    <Prompt description="Product reveal: closed box to open box">
      Start frame is the closed gift box. End frame is the open box with the product visible inside. Generate a smooth 5-second transition where the lid lifts off and the camera tilts down to reveal the contents. Soft studio lighting.
    </Prompt>

    <Prompt description="Before and after: empty room to furnished">
      Start frame is the empty room. End frame is the same room fully furnished. Generate an 8-second morph transition where the furniture fades in piece by piece. Same camera angle throughout.
    </Prompt>
  </Step>
</Steps>

<Note>
  A→B transitions work best when the two images share a similar composition and the change between them is clear: a position shift, a state change, or a visual transformation.
</Note>

## What you cannot do

* You cannot use a video as the source (image only). Use [Edit a Video](/edit-a-video) for video-to-video changes.
* You cannot control exact frame-by-frame motion. The AI interprets your prompt.
* You cannot set a specific frame rate. The model determines this.

## Next steps

<CardGroup cols={2}>
  <Card title="Motion Control" icon="person-walking" href="/motion-control">
    Transfer motion from a reference video onto your image for precise control.
  </Card>

  <Card title="Edit a Video" icon="pen-to-square" href="/edit-a-video">
    Change the style or content of an existing video clip.
  </Card>
</CardGroup>