What each reference type is for
- Images: a character, outfit, product or location that should appear in the video.
- Videos (up to 30 seconds in total): motion, choreography or camera movement to borrow.
- Audio: a voice, line or sound the video should use.
Name your references in the prompt
References are numbered in upload order within each type. Say what each one contributes, so the model does not have to guess.
@Image1 is the main character. She walks through the market in @Image2, copying the camera move from @Video1, and says the line from @Audio1 to the stall owner.
Price in this studio
Without reference videos, Reference to Video costs 5 credits per second at 720p and 8 at 1080p: a 10-second 720p clip is 50 credits.
With reference videos it costs 3 credits per counted second at 720p and 4 at 1080p. Counted seconds are the output length plus the longer of the output length and your reference video time (at most 30 seconds). A 5-second clip with 8 seconds of reference video counts 13 seconds: 39 credits at 720p.
How to make one here
- Choose Kling 4.0 · Reference to Video in the model selector.
- Upload your images, clips and audio — only material you own or have permission to use.
- Write the prompt with @Image1, @Video1 and @Audio1, pick the length, resolution and aspect ratio, and check the quote.