Dreamina

Dreamina is ByteDance's consumer creative studio for AI image and video generation, built alongside CapCut and sharing its ecosystem. Its defining trait is model choice: rather than binding you to a single generator, it routes prompts across several leading models — ByteDance's own Seedream for images and Seedance for video, plus third-party engines including Sora and GPT Image — so you can pick the one that suits a given shot without leaving the canvas.

What it does. Text-to-image generation is the core, with a poster mode that treats typography as a first-class element so text and artwork are composed together rather than pasted on afterwards — useful for social graphics, invitations and ad creative where legible stylised text usually defeats image models. On the video side it animates still images and generates clips from text prompts, and there is a dedicated short-drama format for serialised vertical storytelling. An intelligent reference feature lets you supply your own image to steer composition, colour and overall aesthetic, and the smart editing tools add, remove and vary elements within a finished generation instead of forcing a full re-roll.

Who it suits. Social creators, small e-commerce sellers and marketers producing ad and product visuals who want the output of several frontier models without holding a subscription to each.

Availability and pricing. It runs in the browser and as a free iOS app (iOS 15 or later), first published in August 2026. Generation is metered by credits with paid tiers for higher volumes, so the free allowance is a starting point rather than an unlimited plan. Being a young release, the mobile app's rating is still based on a very small number of reviews.