midjourney-style ai image and video generation, with a --flag prompt language and an api built for agents.
maginary.ai · docs · mcp server · demo
maginary generates images and video from a text prompt using a --flag dsl (--ar 16:9, --v, --flagship, and more) — variations, upscales, pans, and zooms, the same interaction model as midjourney, over a plain async http api.
- mcp —
maginary-mcp(pypi). connect claude desktop, cursor, or any mcp client tomcp.maginary.ai, no install needed. - rest — plain http + openapi, works with any http-capable agent or script.
- x402 — agents can pay per call in usdc on base, no signup. the api answers
402with the exact price; settle it and the response comes back.
- product and dashboard: app.maginary.ai
- api docs: maginary.ai/docs
- parameter reference: maginary.ai/docs/parameters.json
