KinoGenie
Start for free
All toolsAI video generation APIMiniMax H3

Generate AI Video from Text with an API

Turn a text prompt, or an image to animate, into a short video clip with native stereo audio, delivered as an MP4. Run the preset directly on this page or call the same agent-ready endpoint from your product.

Run this tool View API recipe
LIVE TOOL
Run generate video
Describe the clip, pick its length and format, and download the generated MP4 here.
0 / 2000
Try:

Describe the shot, the motion and the sound. MiniMax H3 renders picture and stereo audio together.

Optional. The image becomes the first frame and the clip keeps its aspect ratio; refer to it as <Picture 1> in the prompt.

1376 × 768 px
s

4 to 15 seconds at 24 fps.

5 s at 768p, 6 credits per second. Reserved when the job is queued, refunded if the render fails.

30 credits

Already have an account? Sign in

Useful defaults, typed options.

The tool slug stays stable while your agent supplies named media inputs and a narrow set of documented options.

  • 4-15 s clips at 480p or 768p, with sound
  • Flat price known before the render starts
  • Chain subtitles, watermark or vertical reframe on the result
What this preset does
  1. 01Render the prompt (or animate the image) with MiniMax H3 on a GPU worker
  2. 02Generate picture and stereo audio together
  3. 03Tag the file as AI-generated (EU AI Act art. 50)
  4. 04Return a playable MP4 with a download URL
Stable endpointPOST /api/v1/tools/generate-video
Live API request
{
  "inputs": [],
  "options": { "prompt": "Golden-hour beach, a golden retriever puppy runs toward the camera in slow motion. Audio: soft waves, joyful piano.", "seconds": 5, "resolution": "768p", "aspect_ratio": "16:9" }
}
Successful output example
{
  "result": {
    "output": {
      "filename": "generate-video-output.mp4",
      "contentType": "video/mp4",
      "byteSize": 437021,
      "downloadUrl": "https://cdn.kinopipe.com/…"
    }
  }
}

Know the boundaries before you run.

768p maximum; 4-15 seconds per clip. Wan 2.2 renders 16:9 only, and a request it cannot serve comes back refused with the reason
For image-to-video, pass image_url as the first frame and optionally last_image_url; the clip then follows the image aspect ratio
Rendering takes several minutes, longer at 768p and longer again when no worker is already warm; poll the job or use a webhook
Priced per second of video: 5 credits at 480p, 6 at 768p, reserved when the job is queued
Prompts and results are moderated; no real-person likeness without consent
Two open-weight models: MiniMax H3 by default, or Wan 2.2 TI2V 5B by setting model, which is cheaper and silent

About generate ai video

On MiniMax H3, 30 credits at 768p (6 a second) or 25 at 480p. On Wan 2.2 it is 13 credits, since that model costs 2.5 a second. The price is reserved when the job is queued, so it never changes with render time, and it is refunded if the render fails.

On MiniMax H3 yes, and in the same pass as the picture, so describe the dialogue, sound effects or music you want in the prompt. Wan 2.2 renders no sound at all, which is part of why it costs less. Pick the model accordingly, or add music afterwards with add-music-to-video.

Use the output downloadUrl as the input of any KinoPipe tool (our video editing product): add subtitles, a watermark, reframe to 9:16 or merge it with other clips. Over MCP the agent can chain these calls in one conversation.

MiniMax H3 unless you set model to wan-2-2-5b, which selects Wan 2.2 TI2V 5B. Both have open weights, and both carry an AI-generated tag in the container metadata of every file we return, as required for synthetic media in the EU.

Related media tools

MiniMax H3 API: Text to Video with Native Audio

Call the open-weight MiniMax H3 model over a hosted API: a prompt or a start frame in, a 4 to 15 second MP4 with synchronised stereo audio out, billed per second of video.

Explore