curl-X POST 'https://pollo.ai/api/platform/v1/generation/kling-ai/kling-video-o1/video'\-H'Content-Type: application/json'\-H'x-api-key: YOUR_API_KEY'\-d'{ "input": { "prompt": "A cinematic shot of a golden retriever running through a field of sunflowers at sunset, warm rim light, shallow depth of field", "aspectRatio": "16:9", "duration": 5 }}'
Inkluder eventuelt webhookUrl for at modtage et callback, når opgaven lykkes eller mislykkes.
Valgfrit anmodningsfelt
{"input":{"prompt":"A cinematic shot of a golden retriever running through a field of sunflowers at sunset, warm rim light, shallow depth of field","aspectRatio":"16:9","duration":5},"webhookUrl":"https://example.com/webhooks/pollo"}
Kling Video O1 generates video from a written prompt, reference images, or start-and-end frame pairs in a single request.
With Kling Video O1 on Pollo API, developers add this multi-modal generator to their apps, sending text, image, or reference inputs and choosing aspect ratio and clip length per call.
Key Features of Kling Video O1 API
Multimodal Input Support
The model reads text, a single image, or a set of reference images through one workflow, so a product can switch between idea-first and asset-first generation without extra plumbing.
Multi-Reference Consistency
Up to seven reference images can guide a single generation, letting characters, props, or styles from separate stills appear together in one scene while maintaining feature persistence and identity consistency.
First-and-Last Frame Control
Supply a start image and a tail image, and the model interpolates the motion between them, giving precise control over how a shot begins and where it lands.
Frame-to-Frame Stability
Subjects, lighting, and composition hold steady across the clip, so faces, objects, and backgrounds do not drift or morph between frames during longer actions.
Multi-Step Prompting
Detailed prompts of up to 2,500 characters let you script sequenced actions and staging, so the model can follow a described order of events rather than one static beat.
Flexible Framing
Landscape, portrait, and square outputs are all available, matching the same generation to widescreen players, vertical feeds, or square placements.
Use Cases of Kling Video O1 API
Multi-Character Scenes: Combine reference stills of two or three characters so they share a single shot with consistent faces and outfits.
Reference-Driven Branding: Feed logo, product, and mascot images together to keep brand assets on-model across generated promotional clips.
Storyboard-to-Motion: Turn a first frame and a last frame into an animated shot, useful for previz tools and layout-driven pipelines.
Sequential Action Clips: Script multi-step movements in one prompt, such as a figure that walks in, sits, then reacts to something off-screen.
Illustrated Story Animation: Bring drawn characters from reference art into moving scenes for children's book, comic, or webtoon tooling.
Consistent Product Motion: Animate a product from a single image while keeping its shape, label, and color faithful across the whole clip.
Vertical Content Automation: Generate portrait clips at scale for short-form feeds using text or a single guide image per request.
Concept-to-Video Prototyping: Let teams preview a described shot from a prompt before committing reference assets or production time.
How to Use Kling Video O1 API
Get an API Key: Create a Pollo API account and generate your API key from the developer dashboard.
Choose the Model: Send a POST request to the Kling Video O1 endpoint at /generation/kling-ai/kling-video-o1.
Add Your Inputs: Pass an input object for text-to-video (prompt), image-to-video (image plus optional imageTail), or reference-to-video (images array up to 7), then set aspectRatio and length.
Generate and Retrieve: Submit the task, poll the returned taskId and status, and download the video once processing succeeds.
Prompting Best Practices for Kling Video O1 API
Treat the prompt as a shot list. Name each reference clearly, describe the sequence of actions in order, then set the framing, motion, and mood so the model has a full plan to follow.
A Simple Prompt Formula
Reference or subject setup + ordered actions step by step + camera and framing + setting and lighting + style
What You Should Notice
Label Your References: When using multiple images, describe which subject is which so the model places them correctly in the scene.
Order The Actions: Write events in sequence, since the model can follow multi-step staging rather than a single frozen moment.
Anchor The End Frame: With a tail image, describe how the shot should arrive there so the interpolation feels intentional.
Match Aspect To Placement: Pick landscape, portrait, or square early to fit the target player or feed.
Keep Subjects Distinct: Give each referenced character or object clear traits so consistency holds across the clip.
Example Prompts
Multi-Reference Meetup
"Using the three reference images, the young inventor, the copper robot, and the cluttered workshop, place all three in one scene. The inventor waves, the robot tilts its head in response. Slow dolly-in, warm lamplight, hand-drawn illustration style."
First-to-Last Frame Transition
"Start from the pencil sketch of a lighthouse, end on the finished watercolor version. Color and texture bloom across the frame from left to right as the sky brightens. Static camera, soft daylight, gentle painterly transition."
Sequential Action Beat
"A baker enters frame carrying a tray, sets it on the counter, dusts flour from her hands, then looks up and smiles at a customer. Medium shot, morning window light, warm realistic style, steady handheld camera."
Reference-Faithful Product Motion
"From the single reference image of the ceramic teapot, pour steaming tea into a matching cup while steam curls upward. Keep the glaze pattern and shape exact. Close side-on shot, soft kitchen light, clean realistic style."
Kling Video O1 vs Vidu Q1 vs Runway Gen-4
Capability
Kling Video O1
Vidu Q1
Runway Gen-4
Multi-reference input
✅ Up to 7 reference images
✅ Multiple reference subjects
✅ Reference-guided
First-and-last frame control
✅ Start and tail image
✅
❌ Limited
Text, image, and reference in one model
✅
✅
✅
Multi-step prompt following
✅ Long ordered prompts
✅
✅
Aspect ratio options
16:9, 9:16, 1:1
Multiple
16:9 and vertical
Recommended For
Reference-consistent, sequenced shots
Multi-subject reference scenes
Stylized director-led clips
Why Choose Kling Video O1 API?
Kling Video O1 stands out when a shot must combine several references, hold every subject consistent, and follow an ordered sequence of actions rather than one isolated moment.
Through Pollo API, you reach Kling Video O1 with one API key alongside 300+ leading video and image models, with clean docs, task polling and webhooks.
Start with the Kling Video O1 endpoint on Pollo API to ship reference-driven, multi-step video features, then compare it against Vidu, Runway, and other Kling models without touching your integration.
Kling Video O1 API FAQs
What is Kling Video O1?
Kling Video O1 is a multi-modal video model from Kling AI that generates clips from text, a single image, or up to seven reference images, with strong frame consistency and multi-step prompting.
Does Kling Video O1 API support reference images?
Yes. Reference-to-video accepts up to seven image URLs in one request, so several characters, products, or styles can appear together in a single consistent scene.
Can it control the start and end of a shot?
Yes. In image-to-video you can supply both a starting image and a tail image, and the model interpolates the motion between the two frames.
What aspect ratios and lengths are available?
Kling Video O1 outputs 16:9, 9:16, or 1:1, with clip lengths from 3 to 10 seconds depending on the input mode you choose per request.
How does it handle sequenced actions?
Prompts up to 2,500 characters let you describe events in order, so the model can stage multi-step movement instead of a single static beat.
Why run Kling Video O1 through Pollo API?
Pollo API gives one integration for Kling Video O1 and 300+ other models, with documentation, task tracking and webhooks.
