Guides

    Kling Image O1: change one locked still (3cr)

    Kling Image O1 is an edit. Point at the object. A new text-to-image throws away the still you approved.

    Versely Team3 min read

    Kling Image O1 is an edit. Point at the object. A new text-to-image throws away the still you approved.

    Kling Image O1 sits in the catalog as image-to-image and edit-image. Requires image: true. Credits: 3. Audio: none. Durations: none. Max output resolution: not set. Qualities listed: 1K and 2K. The AI photo editor is the place to run it. The marketing sentence on the row talks about text prompts; the flags on the row still demand a plate.

    Trust the flags, not the slogan

    Quoted description: "Kling Image O1 is an advanced image generation model by Kuaishou that produces high-quality, detailed images from text prompts." That sentence would put you on a generate. The snapshot does not: requires_image is true, categories are image-to-image and edit-image, max reference is 10.

    If you have no approved still, this is the wrong Kling. If you have an approved still and you want a new composition from a paragraph, this is still the wrong Kling — you will be asking an edit row to pretend it is a sampler. Point at the object. Name the change. Keep the rest.

    Aspects run 21:9 through 9:16 plus auto. Match the plate. 3 credits is not a mood-board budget; it is one instruction against one file.

    1K and 2K, not a 4K promise

    There is no 4K on this row. Qualities: 1K, 2K. Do not write "deliver 4K" in the prompt and expect the catalog to grow a tier. If the delivery spec is 4K, pick an edit that lists 4K, or upscale after you like the 2K.

    No audio. No duration. You are not "previewing motion" here. You are changing a still so the next row — the video one — is not animating a bad hand.

    Ten reference slots exist. Use them for the SKU, the face, the room that must survive. Do not dump ten near-duplicates of the same hero; that is how consistency looks like noise.

    Three credits to keep the approval

    A new text-to-image is a new person, a new bottle, a new legal problem. O1 at 3 credits is how you keep the still you already showed the client and fix the one thing they circled.

    The Kling roster is the largest on the site. Image O1 is one line on it. Video O3 is not this page. Best Kling model will tell you when to leave the family. Stay when the job is: locked still in, pointed instruction, 1K or 2K out, 3 credits, no soundtrack. Inpainting is the same idea with a mask; this row is the 3-credit instruction against the whole plate.

    If the still was never locked, do not edit it. Generate elsewhere. Editing a maybe is how you pay 3 credits to keep a maybe.

    FAQ

    Does Kling Image O1 require an upload?

    Yes. Requires image is true. Categories are image-to-image and edit-image. The description mentions text prompts; the row still wants the plate.

    What resolution should I expect?

    Qualities listed are 1K and 2K. There is no max_output_resolution of 4K on this snapshot. Do not promise 4K from this line.

    Why not just regenerate the image?

    Regeneration throws away the approved pixels. This row is 3 credits to change one object on the file you already signed off.

    Can I animate with Image O1?

    No. Durations are empty. Audio is empty. Edit the still, then pick a Kling video row — or any image-to-video row — once the still is actually right.