Kling O1 Omni video model

Kling dropped its VIDEO O1 model on December the 1st 2025. It mixes a bunch of video tools into one setup. You can use text, images, or video clips & the model turns it into short clips, edits stuff, changes scenes, or tweaks styles. Add, remove or modify subjects with simple prompts.

Kling O1 reads your input - like a picture, video, or a few lines of text - and uses that to build or change a video. No need for frame-by-frame edits. You just type stuff like “remove the person in the back” or “change it from day to night” and it handles the rest. The tool sees your prompt and works out the visual changes pixel by pixel.

There are 5 main features.

1. Input Anything. Combines tasks like Text-to-Video, editing, start-end frame control, restyling and more into one engine.
2. Understand Everything. The tool takes in mixed media and figures out what you’re aiming for. It understands characters, props, and places from different angles.
3. All-in-One Reference. Helps keep characters and objects looking the same across shots. Even if the camera moves, it stays consistent.
4. Powerful Combinations. You can mix actions like changing the background and adding characters all in one go.
5. Control the Pace. You pick how long each clip is, between 3 to 10 seconds. Gives you room to shape the story how you like.

It can handle multi-part scenes too. You can toss in multiple characters or objects. It keeps track of them even in busy scenes. 

According to Kling’s own tests, Kling O1 beats others like Veo and Runway in multiple areas, especially when it comes to keeping stuff consistent and editing complex scenes.

Kling Omni Usage Quick Tips

REFERENCE → VIDEO (Omni Reference)

Inputs required:

1. Text prompt
2. “Images” (for overall style/appearance)
3. “Elements” (for specific characters/objects)

Steps:

1. Upload Images that define the style/look of the final video (not characters).
2. Add Elements for actual characters or objects:
   • Upload Frontal image (main view).
   • Upload Reference images (extra angles; optional but recommended).
3. In your prompt, you can reference assets:
   • @image1, @image2 for style/scene images.
   • @element1, @element2 for characters/objects.
4. (Optional) Use a style constraint like “keep in style of image X.”

PRODUCT VIDEO (Reference → Video)

Steps:

1. Add product as Element 1 (frontal shot + reference shot).
2. Add a style image under Images.
3. In prompt, describe camera actions and refer to product with @element1.
4. Set start frame using @image1 if needed.

VIDEO → VIDEO EDIT (Omni Video Edit)

Inputs:

• Base video
• Elements (replacement objects/characters)
• Prompt

Steps:

1. Upload your input video.
2. Add an Element for what should replace something in the video.
3. In prompt, specify the replacement instruction (e.g., “replace the book with @element1”).

Examples supported:

• Replace character/animal
• Change environment
• Day→Night
• Replace reflection in mirror
• Replace background only

VIDEO → VIDEO REFERENCE (Next-Shot Generator)

Steps:

1. Upload a reference video that the model should continue from.
2. Write a prompt describing the next shot.
3. (Optional) Include “keep the style” or continuity details.
4. Generate the next sequential shot.

IMAGE → VIDEO (Start → End Frame)

Inputs:

• Start frame
• End frame
• Prompt

Steps:

1. Upload a Start frame image.
2. (Optional) Upload an End frame image.
3. Add a prompt explaining the transformation or motion.