Kling Image-to-Video
Animate a photo with Kling Video O1 image-to-video. Identity-locked motion, 5–10s clips, 60 credits/s, optional prompt. Keep faces, props and lighting consistent.
Kling Image-to-Video runs Kling's Omni Video O1 model, which is built around identity consistency. Upload a portrait, product or character image, optionally describe the motion, and the model returns a 5- or 10-second clip while keeping the subject's face, props, color palette and lighting close to the source. It is the simpler Kling image-to-video endpoint on Conut.ai: no sound, no resolution switch, and one clear job — animate the photo without redesigning it. Pricing is 60 credits per second, so a 5-second clip costs 300 credits and a 10-second clip costs 600.
- Duration: 5s / 10s
- Price: 60 credits/s
- Prompt: Optional
- Strength: Identity lock
How to use Kling Image-to-Video
Upload a clean still, let identity lock keep the subject recognisable, and use the prompt only for motion you actually want.
- Pick a still with one clear subject Use a sharp photo where the face, product or character fills a reasonable part of the frame. Busy collages and tiny subjects give the model little to preserve.
- Choose 5 or 10 seconds Five seconds covers most social loops at 300 credits. Use ten seconds only when the action needs a second beat; it doubles the cost to 600.
- Write a motion prompt, or leave it empty Keep it to one action, one camera move and one light direction. If you want the most natural result, submit without a prompt and let the model pick.
- Generate and check identity Pause on several frames and compare the face, hands and props with the source. Identity drift usually appears in the middle of the clip, not the first frame.
- Finish the audio in an editor The output is silent. Add music, ambience or voice in an editor, or switch to Kling O3 Pro or WAN 2.5 if you need generated or lip-synced audio.
Settings reference
Kling Image-to-Video exposes only four controls: the source image, an optional prompt, the duration and the price tier that follows from it.
| Setting | Value | Notes |
|---|---|---|
| Input image | 1 image | JPG, JPEG or PNG, at least 300 px per side, up to 10 MB, aspect ratio 1:2.5–2.5:1. |
| Duration | 5s / 10s | Default 5s. There is no custom duration field. |
| Price | 60 credits/s | 5s = 300 credits, 10s = 600 credits. |
| Prompt | Optional | Empty prompt lets the model choose a subtle, natural motion. |
| Aspect ratio | Model default | No ratio selector on this endpoint; the source image guides the frame. |
| Audio | None | Silent video. No sound toggle and no resolution selector. |
What Kling O1 image-to-video does
Kling's Omni Video O1 model treats the source image as the anchor. Its stated strengths are stable character identity across frames, consistent props and scene elements, and maintained color tones and lighting. In practice that makes it a good fit for photos where the person or object must remain recognisable, such as character portraits, product shots, pets and fashion images.
The prompt is optional, which is unusual for image-to-video. You can upload a photo, leave the prompt empty and let the model choose a subtle, natural movement. Adding a short prompt gives you more control over the camera and action, but the model is designed to work without a detailed scene description.
- Keeps faces, props and color palette close to the source image
- Prompt is optional; a short motion phrase is enough
- 5- or 10-second output
- Silent video with no audio track
- Works best with a clean, high-resolution input image
Kling O1 vs Kling O3 Pro vs WAN and Seedance
Kling O1 is the simpler, cheaper Kling image-to-video endpoint. Kling O3 Pro adds longer durations, native sound and multi-shot prompts, but costs more per second and requires Pro or Premium access. The WAN and Seedance models cover different price and duration ranges.
- Choose Kling O1 when identity consistency matters more than resolution or sound.
- Choose Kling O3 Pro when you need 15 seconds or native audio.
- Choose WAN 2.5 when you need a specific voice track and a lower price.
- Choose Seedance 2.5 when you need 4K or clips longer than 15 seconds.
| Model | Duration | Credits | Audio | Resolution | Best for |
|---|---|---|---|---|---|
| Kling O1 I2V | 5s, 10s | 60 /s | No | Model default | Identity-locked photo animation |
| Kling O3 Pro I2V | 3–15s | 80–100 /s | Native sound | Model default | Longer pro shots with sound |
| WAN 2.2 I2V | 5s, 8s | 30–60 /s | No | 480p / 720p | Control on a budget |
| WAN 2.5 I2V | 5s, 10s | 20–40 /s | A/V sync | 480p–1080p | Talking shots |
| Seedance 2.5 I2V | 5–30s | 60–600 /s | Native audio | 480p–4K | Long, high-resolution clips |
Image requirements that actually matter
The endpoint accepts JPG, JPEG and PNG files, at least 300 pixels per side and no more than 10 MB, with an aspect ratio between 1:2.5 and 2.5:1. Those are the hard limits. The practical limits matter more: a high-resolution image with a clear subject, even lighting and a simple background gives the model more detail to preserve.
Avoid heavily cropped faces, motion blur, text overlays and busy collages. If the subject is small in a wide shot, the model has fewer pixels to work with and identity consistency drops. For portraits, a head-and-shoulders crop usually produces a more stable result than a full-body shot with a distant face.
- JPG, JPEG or PNG; at least 300×300 px; 10 MB maximum
- Aspect ratio between 1:2.5 and 2.5:1
- One clear subject with a readable silhouette
- Even lighting and a simple background
- For portraits, keep the face large in frame
Prompting with or without text
If you leave the prompt empty, the model chooses a restrained, natural motion. That is often the best choice for a portrait where you want a subtle head turn or a breeze rather than a staged action. If you want a specific camera move, describe it in one short clause: slow push-in, gentle orbit, camera holds while hair moves.
Keep prompts short. Kling O1 is designed to preserve the image, so a prompt that asks for a new scene, a new outfit or a large change in composition fights the model's main strength. Describe the motion, the camera and the light, not a completely different picture.
- Empty prompt: subtle, natural motion chosen by the model
- Short prompt: one action + one camera move + one light direction
- Good: “gentle breeze, hair moving, slow push-in, warm sunlight”
- Avoid: “change the background to a city and give her a red dress”
- For products, describe the camera orbit and the reflection, not the product itself
Limitations and when to use another model
Kling O1 produces silent video and does not expose a resolution selector, a negative prompt or a seed on this endpoint. Clips are limited to 5 or 10 seconds. At 60 credits per second it is more expensive than the WAN 2.2 and 2.5 models, so it is a poor choice for cheap bulk drafts.
It is a strong choice when the subject must stay recognisable and you do not need sound. If you need dialogue, lip sync or a specific voice track, WAN 2.5 is cheaper. If you need 15 seconds and native sound, Kling O3 Pro is the direct upgrade.
- No audio; add music or voice in an editor
- No resolution selector on this endpoint
- No negative prompt or seed exposed
- Maximum 10 seconds per generation
- Higher per-second cost than WAN 2.2 and WAN 2.5
Prompt examples for Kling Image-to-Video
- Portrait in a breeze: Gentle breeze, hair moving slightly, she blinks and turns her head a few degrees toward the window, slow push-in, warm afternoon light, shallow depth of field. Small, natural motion is where identity lock looks best.
- Product on a turntable: Camera orbits slowly around the product, soft studio highlights sliding across the surface, subtle reflection on the base, no change to the product shape or label. Keep the product unchanged; let the camera do the work.
- Pet portrait: The dog tilts its head, ears moving, eyes tracking toward the camera, soft window light, shallow depth of field, natural fur detail. Pets tolerate small head and ear movement without losing identity.
- Fashion look: Model turns slowly toward the camera, coat moving with the turn, wind lifting the hair, soft overcast light, subtle handheld camera drift. A simple turn keeps the outfit and face consistent.
- Cinematic character: The character looks up as rain starts to fall, water droplets catching the light, camera slowly tilts up, cool blue night lighting, film still. Weather adds motion cues without asking the model to change the scene.
- No prompt: No prompt — let Kling choose a subtle, natural movement from the source image. Useful for portraits where any large action would break the likeness.
Pro tips for Kling O1 Image-to-Video
This endpoint is priced for identity, not for spectacle. Small, controlled motion is where it looks best.
- Let the empty prompt do the work Portraits often look most natural with no prompt at all. The model chooses a restrained head turn or breeze instead of a staged action.
- Keep the source hard-edged A clean silhouette and even lighting give identity lock more to hold onto. Heavy shadows and motion blur in the source become drift in the video.
- Do not ask for a new scene Prompts that change the background, outfit or composition fight the model's main strength. Use Reference-to-Video when the scene itself must change.
- Test at 5 seconds first Half the price and half the frames where identity can drift. Only extend to 10 seconds after the short version holds the subject.
- Product shots need camera motion Describe an orbit, slider or reflection pass rather than a new product state. The label and shape are what the model is preserving.
Frequently asked questions
Is a prompt required for Kling image-to-video?
No. The prompt is optional. Without one, the model chooses a restrained, natural motion from the source image. Adding a short prompt gives you more control over the camera and action.
How much does Kling image-to-video cost?
60 credits per second. A 5-second clip costs 300 credits and a 10-second clip costs 600 credits.
Does Kling O1 image-to-video generate sound?
No. This endpoint returns silent video. Kling O3 Pro image-to-video supports native sound, and WAN 2.5 supports uploaded audio with lip sync.
What image formats and sizes are supported?
JPG, JPEG and PNG, at least 300 pixels per side, up to 10 MB, with an aspect ratio between 1:2.5 and 2.5:1.
Can I choose the output resolution?
Not on this endpoint. Kling O1 image-to-video does not expose a resolution selector; the model returns its default output. For explicit resolution tiers, use WAN or Seedance models.
What is the difference between Kling O1 and Kling O3 Pro?
O3 Pro is the pro-tier model: 3–15 seconds, native sound at 80–100 credits per second, multi-shot prompts and better detail. O1 is simpler, silent, limited to 5 or 10 seconds, and costs 60 credits per second.
Why does Kling keep the face consistent?
The Omni Video O1 model is designed around identity preservation, using the source image as an anchor for facial features, props and lighting. A clean, high-resolution input gives it more information to preserve.
Can I use Kling image-to-video for products?
Yes. It works well for product shots when the camera moves around a static product. Keep the label readable in the source image and avoid prompts that change the product's shape or color.
Compare with other image-to-video models
- Kling O3 Pro Image-to-Video
- WAN 2.2 Image-to-Video
- WAN 2.5 Image-to-Video
- Seedance 2.5 Image-to-Video
- Sora 2 Image-to-Video
- Browse all AI video models
Specifications
- Provider: Wavespeed
- Model type: Image to video
- Credit pricing: 60 credits per second
- Capabilities: new, popular, multi-shot
About Conut.ai
Conut.ai is an AI image, video and audio generation platform with 50+ creative tools. It supports text-to-image, image-to-video, face swap, motion control, relighting, inpainting and video effects through models including Nano Banana Pro, Seedream, GPT Image, Sora 2, Kling, WAN and Seedance. Start free — no credit card required.
Explore more AI tools
- WAN 2.6 Text-to-Video
- WAN 2.6 Reference-to-Video
- HappyHorse 1.0 Text-to-Video
- HappyHorse 1.0 Image-to-Video
- HappyHorse 1.0 Video Edit
- HappyHorse 1.0 Video Extend
- Seedance 2.0 Text-to-Video
- Seedance 2.0 Image-to-Video
- Seedance 2.0 Fast Text-to-Video
- Object Remove
- Smart Crowd Remove
- Image Watermark Remover