Kling 3.0 Omni online: video from reference photos and video editing

What Kling 3.0 Omni can do

Kling 3.0 Omni is the all-in-one version of Kling 3.0. Besides text and frames it takes up to 7 reference photos — a character, a product, an outfit, a location. Refer to them in the prompt as "photo 1", "photo 2" and the model builds the scene keeping their look and details.

Omni also works with your own video (3–10 seconds): edit it with a prompt — new background, outfit or style — or use it as a reference for motion and camera. Clips are 3–15 seconds, up to 4K, with optional sound. Pay per second; failed generations are refunded.

Use cases

• clips with your character or product from photos
• editing an existing video with a prompt
• ad scenes with several objects
• series of clips with the same hero

How to get started

1. Sign in with Telegram, GitHub or email — it takes half a minute.
2. Open the Kling 3.0 Omni page and write a prompt, optionally adding a photo.
3. Check the price on the Create button and run it. The result appears on the page and in your Generations list.
4. Top up when you need to — any amount, no subscription. Prefer code? Create an API key and call the same model over REST.

Prompt tips for Kling 3.0 Omni

• Use sharp reference photos on a plain background and name them in the prompt: "the woman from photo 1 holds the bottle from photo 2".
• When editing a video, describe only the change: "make it evening, add neon signs".
• Fewer references are followed more precisely: 1–3 photos are enough for one character.

The more specific the prompt, the fewer attempts you need. Instead of "a nice picture", say what exactly, where, in what light and in what style.