pretty_name: MiniMax H3 Video Prompts
license: cc-by-4.0
language:
- en
task_categories:
- text-to-video
- image-to-video
tags:
- minimax-h3
- ai-video
- video-generation
- text-to-video
- image-to-video
- prompt-engineering
- video-prompts
size_categories:
- n<1K
configs:
- config_name: default
data_files:
- split: train
path: train.jsonl
MiniMax H3 Video Prompts
A small, curated collection of 50 structured prompts for text-to-video and image-to-video workflows. It covers cinematic scenes, characters, animation, nature, architecture, product shots, food, social video, and fantasy environments.
Use the prompts to create video
Copy a prompt from the dataset, adapt it to your idea, then generate the finished video online.
Create an AI video with MiniMax3.org →
Dataset details
| Split | Rows | Modes |
|---|---|---|
train |
50 | 30 text-to-video, 20 image-to-video |
Each row includes a short brief, a complete prompt, and structured fields for camera, lighting, motion, constraints, aspect ratio, duration, and use case. The data is intentionally small so every prompt is readable, editable, and useful as a starting point.
Fields
| Field | Description |
|---|---|
mode |
text-to-video or image-to-video |
category |
Visual topic or style |
brief |
Short creative direction |
prompt |
Complete copy-ready video prompt |
camera, lighting, motion |
Structured production direction |
constraints |
Consistency and artifact-avoidance guidance |
aspect_ratio, duration_seconds |
Suggested delivery settings |
use_case |
Example viewing context |
Intended use
Use these prompts as inspiration, prompt-engineering examples, or seed material for your own video workflows. They are not benchmark labels, training captions for real videos, or a substitute for a model's official documentation.
Creation and limitations
All rows in v1.0 are original, manually curated text prompts. They do not include personal data, real-person likeness instructions, copyrighted character names, video files, or claims about output quality. Video models can interpret prompts differently; review outputs before publishing or using them commercially.
Community prompt resources for MiniMax H3 video generation. This dataset is not an official MiniMax dataset and does not contain or redistribute MiniMax model weights.
License
The dataset text is released under CC BY 4.0. Attribute this repository when redistributing adapted rows.