EveryGen AI

The video workspace supports text-to-video and image-to-video, with MiniMax H3 Max and H3 Max Turbo offering 5–15 second clips and synchronized native audio. H3 Max supports 480P and 768P, while Turbo also offers 1080P refinement from a native 768P source. The page notes roughly 2.5s backend inference for a 5s 768P clip, with queue and prompt expansion adding to total wait time. The image workspace defaults to GPT Image 2 and supports text and image modes. Public examples let you explore prompts, check source inputs and visible settings, and copy a direction before writing your own. Browser-based tools handle image preparation and text rewriting.

The page frames EveryGen AI around tasks rather than models: a product launch, for example, might need a clean still, a short clip and a description that sounds right. You can describe a new shot with text, or start from a supported image when you already have a visual direction worth developing. Reference-based work is shown in the examples, such as combining character actions and camera language from one video with characters from images and a background image. For stills, the image workspace suits scenes, illustrations and reference-based drafts. The text tools help reshape wording for a different audience, and the site suggests reviewing one meaningful change at a time rather than changing everything at once.

The page states that annual plans cost about 50% less than 12 monthly payments. Generation requests display credits before you submit, and the available resolution, duration, outputs and audio depend on the selected model. No further plan tiers or per-credit rates are listed in the page content.