Hailuo 2.3 Pro Text to Video: Turn Written Scenes into Clips
Last verified: October 19, 2026
Hailuo 2.3 Pro is built to follow direction. MiniMax improved how it responds to motion commands and credits that to better physics understanding and command following. You use this through the Prompt box: write motion as instructions, not mood. Treat Hailuo 2.3 Pro text-to-video generation as directing a shot, and spell out who moves, what changes and where the camera goes.
In text-to-video, the model makes a video from a text description and nothing else. That means the prompt has to carry the subject, the action, the setting, the light and the camera. The form is just as simple: one prompt and an Enhance Prompt switch that lets the model expand what you wrote before it renders. The prompt is your only lever, so it's where your effort pays off.
Explore more text to video
Hailuo 2.3 Pro at a glance
What goes in and what comes back from each run.
Input
A text prompt of up to 2,048 characters
Output
1 video per run
How do you control the camera in Hailuo 2.3 Pro?
Put each camera command right after the action it frames
Sequence moves across the shot instead of grouping them
Switch Enhance Prompt off once your wording is final
How do you make a video with Hailuo 2.3 Pro?
Three steps take you from a written scene to a downloaded clip.
Choosing between a text prompt and a starting image
When to write the shot from scratch and when a first frame should lead.
| Criterion | Hailuo 2.3 Pro | Alternatives | Best For |
|---|---|---|---|
| Starting point | You start from a written prompt alone, so the model creates the subject, setting and framing from your description. | Hailuo 2.3 image-to-video starts from a first-frame image you supply, together with a text prompt. | Write the shot here when it exists only as an idea or you want the model to design the look. Use image-to-video when a specific still, such as an approved product render, has to open the clip. |
| Camera direction | Camera moves go into the prompt as bracketed commands, placed at the beats where they should happen. | MiniMax documents the same bracketed camera commands for Hailuo 2.3 image-to-video, so a move written for one route works in the other. | Camera control doesn't separate the two routes. Let the starting point decide, and use the same camera commands in whichever you pick. |
Common Hailuo 2.3 Pro problems and fixes
The problems that come up most when writing prompts for this model, and what to change.
My prompt was rejected for sensitive content
The camera goes chaotic when I combine several moves
My detailed prompt doesn't fit in the Prompt box
Frequently Asked Questions
Can I upload a reference image to guide the video?
The generator works from text alone. The Prompt box is its input and Enhance Prompt is its one optional setting. Anything you'd normally show with a reference, such as a character's wardrobe, a product's finish or a color palette, has to be written out. Name materials, colors and shapes plainly instead of leaving them to the model.
How much does it cost to generate a Hailuo 2.3 Pro video?
Vidofy calculates the credit cost of the run you've set up and shows it above the Generate button once you're signed in. You see it before you start the run.
Which camera commands does Hailuo 2.3 Pro understand?
MiniMax documents fifteen bracketed commands for text-to-video: [Truck left], [Truck right], [Pan left], [Pan right], [Push in], [Pull out], [Pedestal up], [Pedestal down], [Tilt up], [Tilt down], [Zoom in], [Zoom out], [Shake], [Tracking shot] and [Static shot]. Type them into the Prompt box exactly as written, brackets included.
Does the lighting stay believable when the camera moves?
MiniMax says this about the model directly: even with dynamic camera movement, it gets close to photorealistic results in lighting direction, shadow transitions and color tones. To make use of that, name your light sources and where they come from, such as a lantern on the left or a window glowing blue. Pair them with a moving shot like [Tracking shot], so the light changes across the subject as the camera travels.