What v7 changed
- Style references (--sref) — lock a visual style from any image and apply it across a campaign.
- Character consistency (--cref) — the same character across scenes; the feature that made serial storytelling possible.
- Doubled native resolution — print-ready output without third-party upscalers.
- Text rendering — legible in-image typography, finally usable for posters and covers.
How Sandbox runs it
Sandbox holds an enterprise partnership seat for Midjourney: your generations run through official channels with commercial usage rights, and they're metered in the same credits as text and voice. More importantly, generations chain: write a shot list with GPT-5, storyboard it with Midjourney, animate selected frames with Runway, narrate with ElevenLabs — one brief in, a finished video out, all inside one workspace.
Prompts that work in v7
Why not just use Midjourney on Discord?
Discord works, but it's one model in a chat window with no comparison and no text/voice/video to pair it with. Install Sandbox on Windows and the same Midjourney generations sit next to DALL·E, Imagen and Stable Diffusion in one credit pool — race them on a brief and keep whichever wins.
Midjourney + DALL·E + Imagen + SD3 — one credit pool
Generate with every image model and keep the best. 3 days free.
No card, no email, no signup · 3 days free · Windows 10/11