AI Video Editing API
UnificAlly serves 9 AI video editing models from ByteDance, Alibaba, Kling and Google through one REST API and one API key. Rates start at $0.022/sec (SeeDance 2.0) or $0.11 per video (Gemini Omni Flash 1.1), depending on how each model bills, and you pay per generation from a prepaid balance with no subscription. Every model below uses the same two calls: POST /v1/tasks to start a job and GET /v1/tasks/{task_id} to fetch the result.
All 9 AI video editing models on UnificAlly
| Model | Provider | Starting rate |
|---|---|---|
| Gemini Omni Flash 1.1 API Text, image, and character references with voice presets; 4-10s videos plus conversational editing. | from $0.11 per video | |
| SeeDance 2.5 API ByteDance 4-30s, 480p to 1080p, native audio, first/last frames, and up to 50 multimodal references. | ByteDance | from $0.129/sec |
| SeeDance 2.0 API ByteDance Up to 9 image/video/audio references, T2V or first/last frame, 4–15s, up to 4K on Pro. | ByteDance | from $0.022/sec |
| SeeDance 2.0 Fast API ByteDance Up to 9 references, T2V or first/last frame, 4–15s at 480p or 720p, faster turnaround. | ByteDance | from $0.0372/sec |
| SeeDance 2.0 Mini API ByteDance Up to 9 references, T2V or first/last frame, 4–15s, 480p or 720p, cost-optimized. | ByteDance | from $0.022/sec |
| Kling 3.0 Omni Edit API Kling Edit videos with reference (style guide) or transform mode, plus image and video elements. | Kling | from $0.09/sec |
| Kling O1 Edit API Kling Reasoning video editing with reference or transform mode, up to 4 reference images. | Kling | from $0.09/sec |
| Wan 2.7 Video Edit API Alibaba Edit 2–10s videos with up to 4 reference images at 720P or 1080P, plus audio passthrough. | Alibaba | from $0.07/sec |
| HappyHorse 1.0 Video Edit API Alibaba Edit a 3–60s video by prompt, swap objects from up to 5 reference images, keep the original audio. | Alibaba | from $0.098/sec |
How to call the AI Video Editing API
Every model takes the same request: a model ID and an input object. This example uses kuaishou/kling-o1-video-edit; swap in any model ID from the table and the input fields from its model page.
# 1. Start a task
curl -X POST https://api.unifically.com/v1/tasks \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "kuaishou/kling-o1-video-edit",
"input": {
"prompt": "Transform into anime style",
"video_url": "https://example.com/video.mp4",
"video_mode": "reference",
"mode": "pro"
}
}'
# 2. Read the result with the task_id from step 1
curl https://api.unifically.com/v1/tasks/TASK_ID \
-H "Authorization: Bearer YOUR_API_KEY"AI Video Editing API FAQ
- How much does the AI Video Editing API cost?
- Each model has its own rate, shown in the table above and on the pricing page. The lowest listed rates right now are $0.022/sec (SeeDance 2.0) and $0.11 per video (Gemini Omni Flash 1.1). You pay per generation from a prepaid balance: no subscription, no monthly minimum, and the balance does not expire. Failed generations are not charged, and a request with dry_run set to true returns the exact price without running the job.
- How do I call these AI video editing models from code?
- Send POST https://api.unifically.com/v1/tasks with your API key, a model ID (for example kuaishou/kling-o1-video-edit) and an input object. The response holds a task_id: poll GET /v1/tasks/{task_id} until the status is completed, or pass a callback_url to receive the result by webhook. The request shape is the same for every model; only the model ID and the input fields change.
- How is AI video editing different from video generation?
- An editing model starts from your own footage: you send a video_url and a prompt, and it changes that clip, for example restyling it or replacing a subject with a reference image, instead of generating a new video from nothing. Each model page lists the edits it supports and its input limits.
- Can I try the AI video editing models before topping up?
- Yes. A new UnificAlly account starts with a $0.50 signup balance, and every model page has a browser playground that runs on it, so you can compare outputs before writing any code.