How to Create Video from Image

Animate a still image into video. In Get3W this is the first-to-video run type — your image becomes the first frame.

Quick Start

Video takes longer than images, so submit asynchronously and poll:

cURL

curl -X POST "https://api.get3w.com/v1/kling/kling-3/first-to-video" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "image_url": "https://your-cdn.com/your-image.png",
    "prompt": "A person slowly turning their head and smiling",
    "duration": 5,
    "resolution": "1080p"
  }'

Python

import time
import requests

headers = {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json"
}

task = requests.post(
    "https://api.get3w.com/v1/kling/kling-3/first-to-video",
    headers=headers,
    json={
        "image_url": "https://your-cdn.com/your-image.png",
        "prompt": "A person slowly turning their head and smiling",
        "duration": 5,
        "resolution": "1080p"
    }
).json()

time.sleep(task.get("estimated_duration") or 30)

while True:
    data = requests.get(
        f"https://api.get3w.com/v1/requests/{task['id']}",
        headers=headers
    ).json()

    if data["status"] == "completed":
        print(data["outputs"][0])
        break
    if data["status"] == "failed":
        print(f"Failed ({data['code']}):", data["error"])
        break

    time.sleep(5)

Choosing a Model

Slug	Notes
google/veo-3.1/first-to-video	Highest quality, native audio
google/veo-3.1-fast/first-to-video	Faster and cheaper Veo
kling/kling-3/first-to-video	Strong motion, reliable
kling/kling-3-turbo/first-to-video	Faster Kling
bytedance/seedance-2.5/first-to-video	Good prompt adherence
alibaba/wan-3/first-to-video	General purpose
minimax/hailuo-2.3/first-to-video	Creative, expressive motion
runway/gen-4.5/first-to-video	Cinematic look

Common Parameters

Parameter	Type	Description
image_url	string	URL of the first frame
prompt	string	Description of the desired motion
duration	integer	Video length in seconds
resolution	string	Output resolution, e.g. 720p, 1080p
seed	integer	Seed for reproducibility

Supported durations and resolutions vary per model. Both affect price.

Related Run Types

Run type	Use when
first-to-video	You have one image to animate
first-last-to-video	You have a start and end frame to interpolate between
reference-to-video	You want to guide the video with reference images
text-to-video	You have no source image

Swap the run type in the path and adjust the inputs — for example first-last-to-video takes both a first and last frame.

Workflow

1. Host your image at a publicly reachable URL — the provider fetches it directly
2. Pick a model based on quality, speed, and cost
3. Write a motion prompt describing movement, not just the scene
4. Submit and poll, or pass ?webhook= to get a callback instead
5. Download the output before it expires

Tips

• Describe motion, not content — the image already establishes the scene, so the prompt should say what moves
• Keep motion plausible — motion that contradicts the image tends to produce artifacts
• Start short — validate at 5s before paying for longer clips
• Use webhooks for batches — polling many long video tasks in parallel gets unwieldy fast

Next Steps

• How to Generate Speech — Add narration
• How to Create a Digital Human — Talking avatars
• Webhook Mode — Skip polling