
MiniMax H3 text and image to video

MiniMax H3 text and image to video
Try MiniMax H3 is a browser based AI video generator for creating clips from text prompts, a single image, or a first and optional last frame. It produces videos of 4 to 15 seconds in 768P or 2K resolution, with six aspect ratio options for text to video generation. Prompts can specify the subject, setting, camera, action, timing, editing rhythm, and sound. For frame based generation, the supplied first image guides the visual identity, palette, proportions, and composition, while an optional last frame can define the intended ending. The workflow supports product films, game and interface promotions, narrative sequences, and other motion focused content. Users can direct transitions, visual effects, typography, interface elements, and audio cues, while specifying details that should remain consistent, such as product shapes, labels, faces, or interface states. The site recommends selecting the delivery ratio before generating text based videos, using 768P for iteration and 2K for approved detail, and reviewing the result for identity, text, action order, and sound timing. The task workspace displays prompts and generation settings alongside results and checks active jobs while open. The site provides credit based pricing, with stated costs of 113 credits for a five second 768P video and 183 credits for a five second 2K video. Credit packages have different validity periods, and the displayed package estimates are based on five second videos.