How do I turn a photo into a 10-second Kling 4.0 video?

What can the starting photo control?

Published capability: Kling 4.0 supports image-to-video from a starting image. That image establishes the opening appearance; it does not prescribe every later frame, movement or sound.

Planning advice: Choose a photo with a clear subject and room for the intended movement. Prepare a portrait composition before generation so important details are not near a crop boundary. Describe what changes over time, rather than only repeating what is already visible. A still photo cannot supply a recorded voice or prove how an unseen side of an object should look.

Plan a 10-second, 9:16 shot with stereo sound

Published limits: The official Kling 4.0 release note lists 3–30 seconds, 9:16 among the supported aspect ratios, 720p, 1080p and 4K output, and two-channel stereo audio. A 10-second portrait plan fits those published limits; it is not a guarantee of a successful result.

1. Plan one action: a ceramic cup slowly rotates as steam rises. Keep the background and camera movement simple.
2. On the official platform, select the available image-to-video workflow and set 10 seconds, 9:16 and your desired supported resolution. Check availability there before submitting.
3. Describe the sound separately: quiet room ambience and a subtle left-to-right cloth movement, with no dialogue or music.

An illustrative prompt, not a tested result

This original planning example has not been generated or tested by us:

Use the supplied photo as the opening frame. A ceramic cup rests on a wooden table in soft morning light. Across ten seconds, the cup turns slightly while a thin ribbon of steam rises. Keep its shape, glaze and the table consistent. Hold a steady portrait composition with a very gentle push-in. Add quiet stereo room ambience and a soft cloth rustle moving from left to right. No speech, music or added lettering.

Which frame or reference approach fits?

Single first frame

Start from one photo when the opening matters most; describe the subsequent action in the prompt.

First and last frames

Standard Kling 4.0 supports both endpoints. Use them when the closing composition matters too; intermediate motion still needs review.

Multi-keyframes

Up to 10 keyframe images can mark intermediate beats. Plan these when a start and finish do not explain the intended sequence.

Omni Reference

References guide identity, motion or style rather than simply defining endpoints. The published combined limit is 15 items, including up to 10 images or 5 videos totaling 30 seconds; voice references are voice-only. These are shared limits, not additive allowances.

Those are standard-model capabilities. See Kling 4.0 versus Flash before assuming the same frame controls apply to Flash.

Check the actual output before publishing

• Inspect file duration and dimensions: is it 10 seconds and 9:16?
• Review the subject, edges and motion throughout, not just the thumbnail.
• Check for two audio channels and listen for unwanted sounds, clipping or missing spatial movement. A stereo specification does not guarantee that your requested sound was followed.
• Confirm the downloaded resolution rather than assuming 4K.

Source bounds, checked October 23, 2026: 10-bit HDR at 1080p and 4K, and video extension up to two minutes, are listed as coming soon. Neither is needed or promised for this plan.

Return to the Kling4AI planning workspace to draft your prompt. For actual generation, use the official Kling AI platform.