Kling O3 Standard Image to Video with Reference

Upload an image and a prompt, and Kling O3 Standard animates it. The reference image slot lets you anchor the subject's appearance — useful when the person or object in the output needs to match something specific. Start and end frame inputs give you control over where the clip begins and ends.

Default output is 5 seconds at 24fps, 16:9. Audio generation is optional and runs in the same pass.

How do you use Kling O3 Standard image to video with reference?

Upload a start image, write a prompt describing the action, and Kling O3 generates the animation. The reference image slot anchors subject appearance. End image, multi-prompt, audio, and shot type are all optional — most runs need only the start image and prompt.

Start image The frame the video generates from. Upload a clear, well-composed still. The model animates outward from this frame, so the composition here sets the scene.

Prompt Describe the action you want. The example in the workflow: "make this lady share a sentimental thought." Be specific about expression, motion, and mood. The model follows action descriptions closely.

Reference image Optional. Upload a face or subject reference to anchor who appears in the output. Leave it blank if your start image already shows the subject clearly and you don't need cross-image consistency.

Element 1 frontal image A second reference slot, specifically for a front-facing view of the subject. Use it alongside the reference image when you need tighter face consistency.

End image Optional. Upload a target frame to guide where the clip should land visually. Leave it blank and the model generates the ending from the prompt and start frame alone.

Duration 5 seconds by default. Increase if the action needs more time to play out.

Aspect ratio 16:9 by default. Switch to 9:16 for vertical output or 1:1 for square.

Shot type "Customize" by default, letting your prompt steer the framing. Pick a specific option to lock in a camera angle.

Generate audio Off by default. Turn it on to generate audio alongside the video in the same run.

Multi-prompt Optional. Pass a sequence of prompts to guide different sections of the clip. Useful when the action or mood shifts mid-video.

What is Kling O3 Standard image-to-video with reference good for?

This workflow is for animating a still image when you need the subject's appearance to stay consistent with a separate reference. The start/end frame system gives you more control over the clip's arc than a prompt-only workflow. Good for character-driven content, portrait animation, and scenes where the ending frame matters.

Good scenarios: portrait animation where the subject needs to match a reference photo. Scenes where you want a specific start composition and a specific end state. Content where audio needs to be generated in the same run without a separate step.

If you don't have a reference image and just want to animate a still, the extra inputs don't add friction — leave them blank and treat it as a standard image-to-video run. If you need to edit existing footage rather than generate from a still, use the video-to-video workflows instead.

FAQ

What's the difference between the reference image and the start image in Kling O3? The start image is the first frame of the video — the scene the model animates from. The reference image anchors what the subject looks like across the clip, independent of the starting frame. Use both when you're compositing a subject into a scene they're not already in.

Does Kling O3 Standard support end frame control? Yes. Upload an end image and the model tries to guide the clip toward that final frame. Leave it blank and the ending is generated from the prompt and start frame alone.

What does multi-prompt do in Kling O3 image to video? It lets you pass a sequence of prompts to guide different sections of the clip. Useful for longer videos where the action, expression, or mood changes partway through. For a single continuous action, one prompt is enough.

Can Kling O3 Standard generate audio in the same run? Yes. Toggle generate_audio on and audio is generated alongside the video. Leave it off if you're adding sound in post.

How do you run Kling O3 image to video with reference online? You can run it online through Floyo. No installation, no setup. Open the workflow in your browser, upload your image, and hit run.