Midjourney v7 — what changed and what it means for your workflow
April 1, 2026 · 6 min read
What's new
Midjourney v7 dropped last week with three major changes: native text rendering (finally), coherent multi-image generation, and a web-based editor that replaces the Discord-only workflow.
The Discord requirement was the biggest barrier to adoption. With v7, you can now use Midjourney entirely through midjourney.com — upload reference images, edit generations, and manage your library without touching Discord.
Native text in images
Previous versions couldn't reliably render text — ask for a sign saying 'Open' and you'd get 'Oepn' or 'Opne.' V7 solves this with a dedicated text rendering pipeline. You can now specify exact text, font style (serif, sans, handwritten), and placement.
Why this matters: product mockups, social graphics with headlines, event posters — all the use cases where you'd generate the image in Midjourney and then add text in Canva. Now you can often skip the Canva step entirely.
Multi-image coherence
Ask v7 for 'a product photo of a candle, then the same candle in a lifestyle setting, then the same candle in gift packaging' and you'll get three images of the same candle. Previous versions would generate three different candles.
This is huge for brand consistency. You can now generate an entire product photography set — studio shots, lifestyle, packaging, social — and the product looks the same across all of them.
What this means for your workstation
If you followed our photo-shoot brick, v7 makes it significantly better. The --sref (style reference) flag now works across multiple generations, and the web editor lets you make quick edits without re-generating.
The Midjourney → Canva pipeline we built still works, but v7 reduces how often you need Canva for text overlay. For simple social graphics with a headline, v7 can now handle the whole thing.