maginary.ai reads your mind and generates images/videos using 40+ underlying models it has an excellent midjourney-like syntax and i'm also rolling an easy way for people building chatbots to integrate maginary's powerful image generation/editing via just an mcp-like integration, no complex apis needed. two new models just landed in maginary: Seedance 2.0 and GPT-image-2. no dropdowns, no app-switching. both live in maginary’s premium pool, just add "--flagship" to your prompt

hey ph, i'm Luke (@xucian_). built maginary because keeping up with every new image and video model, ui, and parameter is exhausting. the pitch is simple: one prompt, every model. think openrouter for visual ai. instead of dropdowns, routing happens right in your prompt: plain prompt: maginary picks and mixes fast, cheap curated models --flagship: routes into the premium pool (gpt-image-2, seedance 2.0, nano banana pro, sora 2) direct flag: pin a specific model like --gpt2high or --seedance2pro just added gpt-image-2 (which finally renders real, legible text) and seedance 2.0 (video with native audio and lipsync). there's also a full api and mcp for agents, including an upfront cost-check endpoint so you know credit costs before running anything. pay as you go, no forced subscriptions. you can test sample prompts at app.maginary.ai without signing up. around all day if you have questions or feedback!

this launch is long overdue, i've actually launched maginary in sep 2025, but it was called mjapi, a poor transition from a previous product. it linked to the old site which now sells a different product.

so while the new branding was now set and ready for some time, i figured i'd also launch it here. currently there are 4k users in our (sqlite! haha) database.

maginary.ai

openrouter for image/video with a midjourney-like ux

openrouter for image/video with a midjourney-like ux

Forum Threads

how long have you been using maginary for? feedback so far?

Forum Threads

how long have you been using maginary for? feedback so far?