MiniMax H3 Video: Image-to-Video or Text-to-Video?

If you have a photo you want to bring to life, start with image-to-video. If you have an idea but no starting image, text-to-video is the more direct route. Both can make sense for a MiniMax H3 video—the useful question is what you want to control before generation begins.

I2video provides both workflows with MiniMax H3 Max. The choice changes the starting point: image-to-video begins with an uploaded frame, while text-to-video begins with your written description. Neither guarantees that every detail will stay exactly as prompted, so preview the result before using it.

When to use image-to-video

Choose image-to-video when the source image already contains something you want to keep recognizable: a product, a character, a portrait, or a particular composition. The image establishes the opening look; your prompt can focus on what moves and whether the camera moves.

For example, with a product photo, you might write:

Slowly move the camera closer to the bottle. Keep the bottle upright and centered, preserve its shape and label, and let the background remain still.

This is a useful starting point when you care about the original framing or the subject's appearance. It is not a lock on every pixel: generated frames can still vary, especially around fine details, text, hands, and fast movement. For the upload and generation steps, see the MiniMax H3 Max image-to-video walkthrough.

When to use text-to-video

Choose text-to-video when you are starting from a scene, mood, or action rather than a finished image. You describe the subject, setting, movement, and camera direction, and the model creates the visual starting point as part of the clip.

For example:

A small red paper boat drifts along a rain-filled city gutter at night. The camera tracks beside it at water level as reflections ripple across the pavement. The boat slows beneath a warm streetlight and continues out of frame.

This gives you room to explore an idea without first making or finding a matching image. The trade-off is that you are describing the whole scene instead of anchoring it to a specific source frame. Keep the first prompt focused; adding many subjects, actions, and camera changes at once makes it harder to tell which instruction needs adjustment.

A quick way to choose

| What you have or need | A practical starting point | | --------------------------------------------- | --------------------------------------------- | | A photo whose subject or composition matters | Image-to-video | | A scene idea but no source image | Text-to-video | | A product shot that should stay recognizable | Image-to-video with one restrained motion | | A new setting, character, or visual concept | Text-to-video | | You are unsure what motion fits a still image | Image-to-video, starting with a subtle action |

If both sound plausible, ask yourself: Do I want the video to begin from this exact image? If yes, upload it. If not, describe the scene in text.

Keep the first attempt manageable

Whichever workflow you choose, describe one main action and one camera choice. For an image, say what should change from the uploaded frame. For text, establish the subject and setting before adding movement. Then watch the middle and ending of the clip, not just the first frame.

If the motion feels off, change one part of the prompt and try again. The MiniMax H3 Max prompt guide has more examples for making movement easier to follow, and these camera movement prompts cover common shot directions.

Generation uses credits, and longer clips cost more. Check the credit amount shown in the tool before starting another attempt. You can try image-to-video or text-to-video in I2video and compare which starting point fits your idea.