Kling 4.0 Reference to Video

Kling 4.0 builds a new video from your references: up to 10 images, 5 video clips and 7 characters or objects, kept consistent across a clip of up to 30 seconds.

📄 About Kling 4.0 Reference to Video

Kling 4.0 Reference to Video is built for consistency. Instead of describing a character or product from scratch, you give the model the material you already have - up to 10 reference images, up to 5 short reference videos and up to 7 elements - and refer to them in the prompt as @Image1, @Video1 or @Element1. Kling 4.0 then generates a new clip of 3 to 30 seconds, in 720p or 1080p, in which those people, products, places and styles stay recognisable. Elements are the strongest form of reference: each one bundles a frontal image with up to three extra views of the same character or object from other angles, so the model understands its shape from every side. Reference videos are useful when a motion, a camera move or a setting should be echoed in the new clip. Aspect ratio can be set to Auto, which adapts the frame to your references, or fixed to 16:9, 9:16, 1:1 or 21:9. Multi-shot mode lets you split the result into several shots with their own prompts and durations, which is ideal for keeping one spokesperson or one product consistent across an entire ad sequence. On JAI Portal you pay per second. When you add reference videos, their seconds are billed together with the output at a lower per-second rate (at least twice the output length, as the provider bills it); the exact price is always shown before you generate, and failed generations are refunded.

✨ Key Features

Up to 10 reference images, 5 reference videos and 7 elements in one request

Refer to every input in the prompt as @Image1, @Video1, @Element1

Elements with a frontal view plus up to 3 extra angles for strong consistency

3-30 second videos in 720p or 1080p

Auto aspect ratio that adapts to your references

💡 Use Cases

⚡Keeping one character or spokesperson consistent across a campaign

⚡Showing a real product with its exact packaging in new scenes

⚡Placing people or products into a specific location

⚡Recreating the motion or camera move of a reference clip

🎯 Best For

🎯 Teams who need the same character, product or style to survive across many generated videos - brand campaigns, series content and product marketing.

👍 Pros

✓Strong identity and product consistency from references

✓Images, videos and multi-angle elements in one request

✓Up to 30 seconds per generation in 1080p

✓Auto aspect ratio that matches the references

⚠️ Considerations

△Reference videos are billed together with the output

△Reference videos must be 3-30 seconds each and 30 seconds combined

△Prompts must name the inputs (@Image1, @Element1...) to use them reliably

📚 How to Use Kling 4.0 Reference to Video

Upload reference images and, optionally, reference videos.

For key characters or products, add an element with a frontal view and 1-3 extra angles.

Write the prompt and refer to your inputs as @Image1, @Video1, @Element1.

Pick duration, resolution and aspect ratio, then check the price above the Generate button.

Generate, then download or keep editing the result.

Ready to try Kling 4.0 Reference to Video?

Get 10 free credits — no credit card required

Start Free →

Frequently Asked Questions

Up to 10 images (including the optional start frame), up to 5 videos and up to 7 elements.

An element is one character or object described by a frontal image plus 1-3 extra views from other angles. Refer to it in the prompt as @Element1, @Element2 and so on.

When you add reference videos, the provider bills their seconds together with the output at a lower per-second rate, with a minimum of twice the output length. The final price is shown before you generate.

You pay per second of video with pay-as-you-go credits. 1080p costs about 1.5 times as much as 720p, and the exact price for your settings is shown above the Generate button before you start. Failed generations are refunded automatically.