Video Generation

HiAPI supports multiple video generation models for text-to-video and image-to-video tasks.

Available Models

Model	Provider	Max Duration	Resolution	Credits
Sora	OpenAI	60s	—	100
Kling	Kuaishou	—	—	80
Seedance 1.5 Pro	ByteDance	12s	1080p	Dynamic
Wan 2.7 T2V	Alibaba	15s	1080p	100
Wan 2.7 I2V	Alibaba	15s	1080p	100

Text-to-Video

cURL

curl -X POST https://api.hiapi.ai/v1/videos \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "wan2.7-t2v",
    "prompt": "A Shiba Inu chasing butterflies under cherry blossoms",
    "size": "1920*1080",
    "seconds": 5
  }'

Python

import requests

response = requests.post(
    "https://api.hiapi.ai/v1/videos",
    headers={
        "Authorization": "Bearer YOUR_API_KEY",
        "Content-Type": "application/json"
    },
    json={
        "model": "wan2.7-t2v",
        "prompt": "A Shiba Inu chasing butterflies under cherry blossoms",
        "size": "1920*1080",
        "seconds": 5
    }
)

task_id = response.json()["data"]["task_id"]
print(f"Task created: {task_id}")

Image-to-Video

Animate a still image into a video using Wan 2.7 I2V:

curl -X POST https://api.hiapi.ai/v1/videos \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "wan2.7-i2v",
    "prompt": "The scene comes alive with gentle wind",
    "input_image": "https://example.com/photo.jpg",
    "seconds": 5
  }'

Async Task Pattern

Video generation is asynchronous. The API returns a task_id — poll for the result:

import time
import requests

# 1. Submit task
response = requests.post(
    "https://api.hiapi.ai/v1/videos",
    headers={"Authorization": "Bearer YOUR_API_KEY", "Content-Type": "application/json"},
    json={"model": "wan2.7-t2v", "prompt": "sunset timelapse", "seconds": 5}
)
task_id = response.json()["data"]["task_id"]

# 2. Poll for result
while True:
    result = requests.get(
        f"https://api.hiapi.ai/v1/videos/{task_id}",
        headers={"Authorization": "Bearer YOUR_API_KEY"}
    ).json()
    if result["data"]["status"] == "completed":
        print("Video URL:", result["data"]["url"])
        break
    time.sleep(5)

Video Parameters

Parameter	Values	Models
seconds	3, 5, 8, 10, 12, 15	Wan 2.7, Seedance
size	1920*1080, 1080*1920, 1280*720	Wan 2.7
duration	5, 10	Sora, Kling