Google AI’s Veo 2 model generates high-quality videos from images with realistic spatial placement and motion, leveraging Google’s Gemini and Imagen models.
curl-X POST 'https://pollo.ai/api/platform/v1/generation/google/veo-2/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, "resolution": "720p" }}'
Optionally include webhookUrl to receive a callback when the task succeeds or fails.
Optional request field
{"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,"resolution":"720p"},"webhookUrl":"https://example.com/webhooks/pollo"}
Schema
Input fields for the selected endpoint and mode.
Field
Type
Required
Description
prompt
string
Yes
Text prompt describing the content to generate.
aspectRatio
string
No
Output aspect ratio, as width:height (e.g. 16:9).
duration
number
No
The duration of the generated video, in seconds.
resolution
string
No
Output resolution of the generated media (e.g. 720p, 1080p, 2K, 4K). Values are case-insensitive.
Veo 2 is strong when motion needs to feel physically grounded: poured liquid splashes, dropped objects settle, and moving subjects carry weight instead of sliding unnaturally through the frame.
With Veo 2 on Pollo API, developers can add text-to-video and image-to-video generation to apps that need short 720P clips in 5 or 8 seconds, with portrait and landscape output options.
Key Features of Veo 2 API
Real-World Physics
Veo 2 models gravity, momentum, and material behavior, so falling objects, flowing water, and swinging limbs move in ways viewers can recognize from real life.
Natural Motion Flow
Camera drift, subject movement, and background details progress smoothly across the clip, reducing the stutter and warping that can make short generations feel artificial.
Prompt Comprehension
Detailed descriptions of subjects, actions, and staging remain visible in the generated clip, helping complex scene instructions produce more specific results.
Text and Image Guidance
Start from a written brief for a new scene, or supply a source image to animate an existing frame while preserving its subject and composition.
Clean Detail at 720P
Textures, edges, and lighting stay readable across the frame, making Veo 2 useful for short social clips, previews, and product visuals.
Use Cases of Veo 2 API
Nature and Wildlife Clips: Generate short shots of animals moving, water flowing, or foliage reacting to wind, where believable motion carries the scene.
Physics-Driven Explainers: Show liquids pouring, objects toppling, or gears turning with motion grounded enough for science and how-things-work content.
Image Animation Tools: Turn a still frame into a moving shot for apps that convert photos, illustrations, or concept frames into short videos.
Food and Recipe Motion: Produce close shots of batter pouring, steam rising, or ingredients folding, where texture and flow matter most.
Short Social Video: Create quick portrait or landscape clips for feeds, stories, thumbnails, and looping visual backgrounds.
Concept Previsualization: Let creative teams preview a described shot before production, testing motion, framing, and scene behavior early.
Automated Content Pipelines: Generate short clips from structured prompts for platforms that need consistent, motion-aware video at scale.
How to Use Veo 2 API
Get an API Key: Create a Pollo API account and generate your API key from the developer dashboard.
Choose the Model: Send your request to the Veo 2 endpoint at /generation/google/veo2.
Add Your Inputs: Provide a prompt for text-to-video, or an image URL for image-to-video, then set length (5 or 8), aspectRatio (16:9 or 9:16), and resolution (720p).
Generate and Retrieve: Submit the task, check the returned task status, and download the finished video once processing succeeds.
Prompting Best Practices for Veo 2 API
Describe the scene like a physical event. Name the subject, the exact motion, and the force behind it, then add camera direction, setting, lighting, and material details.
A Simple Prompt Formula
Subject + specific motion with physical cause + camera angle or movement + setting + lighting and material style
What You Should Notice
Anchor the Physics: State how something moves, such as splashing, rolling, tipping, bouncing, or settling, so the model has a concrete motion target.
Keep One Main Action: Focus each 5 or 8 second clip on one event so the motion has enough time to read clearly.
Direct the Camera: Name the angle and movement, such as overhead, side-on, locked-off, or slow push-in, to control how the action is framed.
Set the Aspect Early: Choose 16:9 for wide scenes or 9:16 for feed-first clips so composition fits the target format.
Describe Light and Material: Mention surfaces, reflections, softness, or texture so physical details render with more believable weight.
Example Prompts
Flowing Water Scene
"A mountain stream tumbles over mossy rocks, water splitting into spray as it hits a boulder, droplets catching morning light. Slow overhead drift, soft natural lighting, realistic style focused on fluid motion."
Everyday Physics Moment
"A stack of wooden blocks topples across a table, each block tumbling and settling with a small bounce, dust lifting on impact. Side-on camera at desk level, warm indoor light, clean realistic look."
Food Preparation Close-Up
"Thick pancake batter pours from a bowl into a hot pan, spreading outward as tiny bubbles rise to the surface. Tight overhead shot, bright kitchen lighting, glossy realistic texture, gentle steam drifting up."
Animated Still Frame
"Animate this landscape photo: tall grass sways in a slow breeze, clouds inch across the sky, and shadows shift over the field while the framing stays fixed. Subtle natural motion, soft afternoon light."
Veo 2 vs Kling 1.6 vs Runway Gen-3 Alpha
Capability
Veo 2
Kling 1.6
Runway Gen-3 Alpha
Physics realism
Strong material and motion behavior
Good motion handling
Smooth, more stylized motion
Text-to-video
Supported
Supported
Supported
Image-to-video
Supported
Supported
Supported
Prompt comprehension
Strong fit for detailed physical scenes
Good fit for motion-rich prompts
Good fit for directed creative shots
Output focus
Realistic short clips with grounded motion
Dynamic movement and longer scene energy
Stylized visual direction and polished shots
Recommended For
Physics-aware realistic clips
Motion-heavy video concepts
Creative direction and stylized video work
Why Choose Veo 2 API?
Veo 2 is a strong choice when short clips need believable physical behavior, from flowing water and falling objects to animated stills with grounded motion.
Through Pollo API, teams can run Veo 2 alongside a broader image and video model stack from one integration, with clean docs and task tracking built into the workflow.
Start with Veo 2 on Pollo API to build physics-aware short video features, then compare it with other models as your product needs change.
Veo 2 API FAQs
What is Veo 2?
Veo 2 is Google DeepMind's video generation model for realistic short clips with natural motion, physical plausibility, and support for text-to-video and image-to-video workflows.
Does Veo 2 API support text and image inputs?
Yes. You can generate a new scene from a text prompt, or provide an image URL to animate an existing frame while preserving its composition.
What resolution and lengths are available?
Veo 2 on Pollo API outputs 720P clips at either 5 or 8 seconds, with 16:9 landscape and 9:16 portrait aspect ratios.
How well does Veo 2 handle motion?
Veo 2 is especially useful for physical motion, including gravity, momentum, liquid flow, object impact, and grounded subject movement.
Can Veo 2 animate a photo?
Yes. Provide a source image and Veo 2 can add motion to it, turning a still frame into a short video while keeping the original subject recognizable.
Why run Veo 2 through Pollo API?
Pollo API helps teams add Veo 2 without maintaining separate provider integrations, while keeping documentation, task tracking, and model switching in one workflow.
