model: kling-v3-motion. Supply appearance and motion references, submit one asynchronous request, and track the returned img_uuid by polling or callback.model: kling-v3-motion. Supply appearance and motion references, submit one asynchronous request, and track the returned img_uuid by polling or callback.
image_url for a clear upper- or full-body character image, or use at most one Kling element through element_ids. Element input forces character_orientation: video.
video_url. The output follows usable continuous motion from that file; this endpoint has no duration parameter.
https://api.goenhance.ai/api/v1/videos/generations with model: kling-v3-motion.prompt of up to 2,500 characters to describe the character, clothing, scene, or relevant generation direction.image_url may be JPG, JPEG, or PNG, up to 50 MB, at least 300 px per side, with an aspect ratio from 1:2.5 to 2.5:1.video_url must be MP4 or MOV, no larger than 100 MB, with dimensions from 340 to 3,850 px.character_orientation: video accepts 3–30 seconds; image accepts 3–10 seconds. The documented default is video.generate_audio keeps the motion video's original sound and defaults to false.video_url before deducting tokens, so changing the remote file after measurement does not change the billed input.duration; the endpoint rejects it, and output length follows the usable motion extracted from the reference video.Authorization header.model: kling-v3-motion, then save the returned img_uuid before handing work to a background processor.GET /api/v1/jobs/detail or supply a publicly accessible HTTPS custom_callback_url.Submit through GoEnhance at POST https://api.goenhance.ai/api/v1/videos/generations with model: kling-v3-motion. The response returns an img_uuid.
720p costs 6.3 tokens/s ($0.126/s), and 1080p costs 8.4 tokens/s ($0.168/s). GoEnhance measures the motion-video duration before deduction as the request's upper-bound charge.
No. This endpoint rejects duration. The reference video must satisfy the orientation-specific limits, and the output follows the usable continuous motion extracted from it.
video follows the facing direction in the motion video and permits 3–30 seconds. image follows the appearance image and permits 3–10 seconds.
No. generate_audio only keeps the motion video's original soundtrack. It defaults to false; use audio only when you have the necessary rights.
The V3 route accepts at most one value through element_ids (or the single-value alias element_id). Element input forces character_orientation: video.
Supply a public HTTPS custom_callback_url. The callback body matches the job-detail response; non-200 deliveries may be retried up to three times with a three-second timeout.
No. Kling is associated with Kuaishou, but this page describes independent access through GoEnhance. GoEnhance is not presented as Kuaishou or Kling's official API, partner, or model developer.