Text-to-Video

Generate high-quality videos from text prompts using MuAPI's curated selection of leading T2V models.

Quick Start

POST /api/v1/kling-master
x-api-key: YOUR_API_KEY
Content-Type: application/json

{
  "prompt": "A golden retriever running on a sunny beach, slow motion",
  "duration": 5
}

Returns a request_id. Poll for the result:

GET /api/v1/predictions/{request_id}/result
x-api-key: YOUR_API_KEY

Available Models

Model	Speed	Quality	Best For
`veo3`	Slow	Flagship	Photorealistic, cinematic scenes
`veo3-fast`	Fast	High	Rapid prototyping with Veo quality
`kling-master`	Medium	Flagship	Cinematic motion, complex scenes
`kling-pro`	Medium	High	Balanced quality and cost
`kling-std`	Fast	Good	Quick generations
`wan2.1` / `wan2.2`	Fast	High	Open-source quality
`seedance-pro`	Medium	High	Character-consistent video
`seedance-lite`	Fast	Good	Fast Seedance generations
`hunyuan`	Medium	High	Hunyuan video quality
`runway`	Medium	High	Creative video
`minimax-hailuo-02-pro`	Medium	High	MiniMax flagship

Prompt Tips

Be specific about camera movement: "slow dolly push-in", "aerial drone shot"
Describe lighting: "golden hour", "dramatic side lighting"
Include motion details: "in slow motion", "time-lapse"
Specify style: "cinematic", "documentary", "anime"

Parameters

Most T2V models accept:

Parameter	Type	Description
`prompt`	string	Text description of the video
`duration`	int	Length in seconds (model-dependent, typically 5–10)
`aspect_ratio`	string	e.g. `"16:9"`, `"9:16"`, `"1:1"`
`negative_prompt`	string	What to avoid in the output

Check the specific model's playground page for all supported parameters.

Via CLI

muapi video generate "A timelapse of storm clouds over a city skyline" --model veo3-fast