Turn a prompt into a 4–15 second video with native audio. Choose 768P or 2K.

Loading video settings…

From idea to video

Turn a written scene into a MiniMax H3 video

Use MiniMax H3 on Veo 3 Flow to create a short video from a text prompt, with native audio in the same generation. Describe a subject, one clear action and the sound of the scene, then choose the length and format for your first draft.

What you can create with MiniMax H3

  • A scene with its own sound

    Describe ambient sounds, sound effects or spoken dialogue alongside the visuals. H3 generates native audio as part of the video; there is no separate audio toggle on this page.

  • Short clips, with room for an idea

    Choose any whole-second duration from 4 to 15 seconds. Try a brief product reveal, an atmospheric establishing shot or a character moment with one clear action.

  • A format that fits your project

    Choose 768P or 2K and one of six aspect ratios: 21:9, 16:9, 4:3, 1:1, 3:4 or 9:16. Use a wide frame for a landscape or a vertical frame for a mobile video draft.

How to generate a video with MiniMax H3

  1. Describe one scene

    Write who or what is in the frame, what happens, and how the camera moves. Add lighting and sound details. Put any spoken dialogue in quotation marks.

  2. Choose settings and check credits

    Set the duration, resolution and aspect ratio. The generator shows the credit cost before you submit. Sign in and make sure your balance covers the selected settings.

  3. Generate, review and download

    Submit your prompt and wait for the result. Play the completed video to review its picture and sound, then download it. Your generation history lets you return to the task.

MiniMax H3 prompt examples

Use these text examples as a starting point, then change the subject or setting. These are sample prompts, not previews of generated results. Keep the action simple when choosing a short duration.

An atmospheric landscape

A small wooden boat drifts across a still mountain lake at sunrise. The camera slowly moves forward just above the water. Pale mist floats between the pine trees. Soft water ripples and distant birdsong; no dialogue.

A product detail shot

A ceramic coffee cup sits on a dark wooden table beside a window. A hand gently sets a saucer down next to it as steam rises from the coffee. Slow camera push-in, warm morning light, close detail. A soft ceramic clink and quiet room ambience.

A character moment

A traveler in a yellow raincoat stops under a streetlamp on a rainy evening. The camera holds a medium shot as she looks toward the light and says, "We made it." Gentle rain, footsteps fading away, and a calm voice.

Frequently asked questions

What is MiniMax H3 text to video?

This tool uses MiniMax H3 to generate short videos from written prompts, including native audio. It is a text-to-video workflow: describe the scene you want, choose your video settings, and submit a generation task.

Which settings can I choose?

Generate a 4–15 second video at 768P or 2K. Choose landscape, portrait, square or one of the other available aspect ratios. Your credit cost updates with the settings.

Can I describe speech or sound?

Yes. Include dialogue in quotation marks and describe the sound you want in your prompt. H3 generates native audio; this text-to-video workflow has no separate audio switch.

How much does a MiniMax H3 video cost?

The credit cost depends on your selected duration and resolution. Check the live estimate beside the Generate button before submitting. Credits are reserved when the task is accepted; a successful task is not charged a second time.

What happens if generation fails?

Credits are reserved when your task is accepted and refunded if it fails. A temporary connection problem does not mean the task failed. Check its status or return to your generation history.

Can I upload an image here?

This page creates videos from text. Image and reference workflows are separate capabilities and are not available on this page.

Why is my video still processing?

Generation time can vary. A result may also take time to reach your history after the video has finished. If the page loses its connection, check the existing task again instead of submitting the same request twice.

How can I improve a prompt?

Start with one subject and one visible action, then add camera movement, lighting and sound. For a short clip, avoid asking for several scene changes. Review the first result and adjust a specific detail for your next attempt; outputs can vary.

Explore more video tools

Have a scene in mind?

Back to the generator ↑