Video Generationtextimage

Happy Horse 1.1

Alibaba ATH Happy Horse 1.1 视频生成模型，经 Magnific 接入。支持文生视频、首帧图生视频和图片参考生视频，使用 Token360 标准视频字段。

输入

生成方式

仅使用文本描述生成视频，无需上传图片。

提示词

参考图（可选）

支持 PNG、JPG、WebP · 可将文件拖到此区域

Duration

Watermark

Resolution

Aspect ratio

准备开始生成

API 参考

POSThttps://api.token360.ai/v1/videos

# Step 1: Create video generation task
curl https://api.token360.ai/v1/videos \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-api-key" \
  -d '{
  "model": "happy-horse-1.1",
  "prompt": "A cat walking across a room",
  "duration": 5,
  "watermark": true,
  "resolution": "1080p",
  "aspect_ratio": "16:9"
}'

# Step 2: Poll for completion (replace VIDEO_ID with the id from step 1)
curl https://api.token360.ai/v1/videos/VIDEO_ID \
  -H "Authorization: Bearer your-api-key"

API 文档

Overview#

Happy Horse 1.1 is exposed through Token360 POST /v1/videos using normalized video request fields. Use prompt-only requests for text-to-video, frame_images with one first_frame image for image-to-video, and input_references image_url entries for reference-to-video.

Canonical request fields#

Use model, prompt, duration, resolution, aspect_ratio, watermark, seed, callback_url, frame_images, and input_references. Do not send Magnific upstream-only image_urls in Token360 requests.

Submit response fields#

The submit response follows the admin API reference for POST /v1/videos: id, object, status, model, prompt, created_at, parameters, content, response, error, and usage. Poll GET /v1/videos/{video_id} for updated_at, url, video_url, duration, progress, provider_task_id, and terminal error details.