DX Builder
KUAISHOU × DXBUILDER

KLING 4.0 · REFERENCE TO VIDEO

Your references. One cinematic idea.

Bring a face, a place and a movement together. Direct 3–30 second videos with Kling 4.0, in 720p or 1080p.

3–30sOutput duration
720p / 1080pResolution
10 + 5Image references + Video references
8 000Scene direction

01 / REFERENCE TO VIDEO

From intention to a moving scene.

Define the look

01 · Assign each reference a role. Use @image_1 for identity, @image_2 for the setting and @video_1 for movement.

Direct the movement

02 · Describe one clear action, its progression and its ending. Say what must remain unchanged.

Keep the atmosphere

03 · Choose duration and resolution. Check the total credits before sending, then review the result in your library.

02 / @image_1 + @image_2

Scene direction

Use @image_1 for the product and @image_2 for the setting. A slow lateral camera move reveals the product on the table. Preserve its shape, colours and label. End with the product centred and fully visible. Soft ambient room sound, no speech.

03 / DX CREDITS

See the whole price first.

Without a reference video, only output seconds count. With video, billable seconds = output + the greater of input and output. All input clips are added together. Credits are rounded up once, on the total. Optional tools have their own prices.

214credits10s · 720p

04 / REFERENCE TO VIDEO

Know your creative space.

Up to 10 JPEG/PNG/WebP images and 5 reference videos. Each file: 10 MB maximum. Videos: 2–30 seconds each, 30 seconds combined. Prompt: 8,000 characters. This studio uses Reference to Video; keyframes, voice references and character elements are not enabled here. 4K is not offered.

Before you press generate.

Does a reference guarantee an identical result?

No. References guide identity, setting and movement, but details can vary. Review faces, text, clothing, sound and the ending.

Why does a short reference video increase the total?

Video input is billable. For 10 seconds of output and 5 seconds of input, the billed duration is 20 seconds. With 15 seconds of input it is 25 seconds.

Can I control the first and last frames?

This integration is Reference to Video. A reference image is not a guaranteed first or last frame. The separate keyframe mode is not enabled.