Direct scenes, action and sound with text, choosing framing for landscape, portrait or square clips.
Provider: MiniMax
Category: video generation
Endpoint: /v1/tasks
Status: Available
Cost: See live page pricing
Direct scenes, action and sound with text, choosing framing for landscape, portrait or square clips.
Pricing
Standard Usage
Range: 2 - 15
Replace the output soundtrack with audio of at least 2 seconds and at most 15 MB. Longer audio is trimmed; shorter audio is padded with silence. Video duration stays unchanged
Uploaded audio
disabled skips expansion; balanced is the default; quality adds detail and may take longer
Ready for generation
Configure your parameters and click "Run" to see the output here.
View portrait subject framing and a wider environmental shot, then consider water movement, bridge light and soundtrack choices for your own clip.
Portrait framing brings the red boat and stone bridge reflections into a vertical composition, paired with the supplied target audio for a cutaway clip.
Prompt
A red paper boat drifts beneath a stone bridge after rain. Smooth tracking camera, natural ripples, soft birdsong. No text or subtitles.
Target audio input
A 4:3 composition leaves room for the bridge, red boat and water, with small ripples unfolding across an environmental shot or opener.
Prompt
A red paper boat drifts beneath a stone bridge after rain. Smooth tracking camera, natural ripples, soft birdsong. No text or subtitles.
A square frame gathers the ceramic cup, wooden table and rainy window into one quiet composition as the camera moves gently closer. The sound prompt pairs soft rain with a single quiet piano note, giving a lifestyle sequence room to pause.
Prompt
A ceramic cup of hot tea on a wooden table beside a rain-streaked window, steam curling upward, slow push-in, soft rain and a single quiet piano note. No text or subtitles.
The wide frame begins close behind the cyclist, then rises to reveal the stone plaza, long shadows and distant skyline. Birdsong and distant bells in the sound prompt make this a useful opening for a travel or city film.
Prompt
A lone cyclist crosses a wide empty plaza at dawn, long shadows on the stone, the camera rising slowly behind them, birdsong and distant bells. No text or subtitles.
The portrait composition follows the height of the clay vessel as wet hands move upward from its belly to shape a neck, with window light catching the turning ridges and fingertips. The sound prompt asks for a soft wheel whirr and water, adding a sense of everyday studio activity to this close view of the craft.
Prompt
A potter shapes a clay vase on a spinning wheel, wet hands smoothing the curve upward, warm window light across the studio, soft whirring of the wheel and the sound of water. No text or subtitles.
Stay here when all you have is a written idea. If you already have a finished picture, look at the Image-to-Video entry.
Current price: View live pricing
Describe the subject, setting, action, camera direction and sound.
Choose framing and resolution for the destination. Request at least 2 seconds when using target audio.
Review the clip and sound online, then use the request below to create in your app.
curl --request POST 'https://api.hiapi.ai/v1/tasks' \
--header 'Authorization: Bearer YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"model": "minimax-h3-max/text-to-video",
"input": {
"prompt": "A red paper boat drifts beneath a stone bridge after rain. Smooth tracking camera, natural ripples, soft birdsong. No text or subtitles.",
"resolution": "768P",
"duration": 2,
"aspect_ratio": "9:16",
"target_audio_url": "https://static.hiapi.ai/model-onboarding/minimax-h3-max/2026/10/13/1791862423253-b835317bfd92f652-target-audio.wav"
}
}'MINIMAX / H3 MAX · TEXT-TO-VIDEO
Text is all you need. Pick one of six frame shapes from 21:9 to 9:16 and the model arranges subject, action, camera and sound from your description. Already have music or a voice-over? Upload it as target audio to replace the soundtrack.
MODEL OVERVIEW
Text-to-video works when the description makes one shot clear. Put the subject, setting, action, camera move and sound in one passage and the model has enough to arrange the picture — the shortest path from idea to clip.
The frame-shape option sets the canvas, so one description can produce landscape, vertical and square versions. With target audio the clip uses your sound while motion and camera still follow the text. Prompts can run to 50,000 characters.
Every clip was generated from text alone; expand to read its prompt.
The vertical frame stacks the red boat, the arch and its reflection into one column, leaving the subject room to move so the action reads instantly on a phone.
This clip used uploaded target audio, so its soundtrack is that recording. When the music for an edit is already set, you can start from the sound and generate the picture to match.
A red paper boat drifts beneath a stone bridge after rain. Smooth tracking camera, natural ripples, soft birdsong. No text or subtitles.
A 4:3 frame holds the bridge, the boat and the bank together. Small ripples unfold along the camera move, and the action does not need to fill the frame.
It works as an opener for brand stories or place introductions. Light, spatial layout and a slow camera build mood better than a frame crowded with action.
A red paper boat drifts beneath a stone bridge after rain. Smooth tracking camera, natural ripples, soft birdsong. No text or subtitles.
A 1:1 square gathers the ceramic cup, the wooden table and the rainy window, and a slow push-in leaves time to watch the steam rise.
The sound is written into the prompt: soft rain and one quiet piano note. Use it as a pause in a lifestyle video or as a square social post.
A ceramic cup of hot tea on a wooden table beside a rain-streaked window, steam curling upward, slow push-in, soft rain and a single quiet piano note. No text or subtitles.
The wide frame starts close to the cyclist, then rises to reveal the stone plaza, long shadows and the distant skyline.
Birdsong and a distant bell are written into the sound description. Moving from close to wide introduces both person and place in one shot, which suits travel or city openers.
A lone cyclist crosses a wide empty plaza at dawn, long shadows on the stone, the camera rising slowly behind them, birdsong and distant bells. No text or subtitles.
The vertical frame follows the height of the clay as wet hands draw the neck up from the body, with window light on the turning ridges and fingertips.
The soft hum of the wheel and running water are written into the sound prompt, giving the close-up a working-studio feel for craft brands and vertical tutorials.
A potter shapes a clay vase on a spinning wheel, wet hands smoothing the curve upward, warm window light across the studio, soft whirring of the wheel and the sound of water. No text or subtitles.
Go straight from an action idea to a 9:16 shot for short video, story transitions and mobile placements.
Keep the prompt and seed, change only the frame shape, and get landscape, vertical and square versions for each channel.
Upload music or a voice-over as target audio, describe the camera rhythm, and the clip keeps your sound.
Prepare wide establishing footage for article videos and brand stories by writing place, weather and light into the prompt.
Write a scene in the Playground, pick a frame shape and try a clip, then send the same settings through the API. Every version stays in task history so you can compare shapes and soundtracks side by side.
No. This entry creates video from text and does not accept first- or last-frame images. Choose MiniMax H3 Max Image-to-Video when an image should set the opening or ending.
Choose 21:9, 16:9, 4:3, 1:1, 3:4 or 9:16. Landscape suits environments and openers, portrait suits short videos, and square fits feeds. Describe the subject’s position and room for movement so the shot fits the chosen frame.
Start with the subject and setting, then add one continuous action, camera direction and sound. Avoid crowding a short clip with many events. Use prompt expansion for a fuller description and refine it after reviewing the clip.
Yes. This entry generates directly in the chosen frame shape, and our 768P test returned 768×1344.
No. With target audio the video must be at least 2 seconds and the audio itself at least 2 seconds. For clips under a second, let the model generate its own sound.
Keep the prompt and seed fixed and change only aspect_ratio. The same seed does not guarantee identical pictures, but style, color and motion tendencies usually stay closer, which makes selection easier.
Use balanced or quality when you only have a short idea and want the model to add detail. Use disabled when the shot, light and action are already written and you want your own wording kept. Expansion does not change the requested length.
Video is generated in frames, so the final length lands on the nearest frame boundary. In our tests the difference was under a third of a second. For exact edit points, use the length of the returned video.
1080P is an output tier refined from a 768P source and suits final delivery or large-screen playback. For drafting and comparing ideas, 480P or 768P is quicker.
HiAPI Blog
