Midjourney rolls out a unified web editor with a precise inpainting brush

Midjourney just made editing AI images on the web look like a real creative suite.

For months, using Midjourney meant wrestling with Discord commands or dealing with a barebones web interface where every edit lived in a separate dropdown. Now, the company has rolled out a unified web editor that brings all its manipulation tools into one canvas.

The feature is available to anyone who has generated at least 10 images on the platform, accessible at midjourney.com/imagine. While the site remains in alpha, this is the company's biggest move yet to break free from chat-based creation.

Ditching the lasso for a real brush

The biggest workflow upgrade is a new virtual inpainting brush. Previously, users had to rely on clunky rectangular selection boxes and lasso tools to specify regions they wanted to tweak. The new brush lets you paint directly over the pixels you want to change, offering significantly tighter control.

Crucially, all the scattered features are now living under one roof. Canvas extension, directional panning, zooming, and prompt-guided repainting happen in a single, fluid workspace instead of separate menus.

Midjourney CEO David Holz called the update a huge step forward for seamless editing. Alongside the editor, the company rolled out message mirroring between Discord and its web rooms, ensuring active channels like prompt-craft, general-1, and a superuser room for members with over 1,000 generations stay in sync across both platforms.

The pressure is mounting

This interface overhaul comes at a high-stakes moment for the startup.

Open-source challenger Flux.1, recently leveraged by Elon Musk's Grok-2, has proven that top-tier visual fidelity and strict prompt adherence are no longer Midjourney's exclusive domain. While Midjourney shipped v6.1 in late July with a 25 percent speed bump and cleaner anatomy rendering, v7 is still nowhere to be seen.

Meanwhile, the legal heat is rising. Just last week, a federal judge rejected bids by Midjourney and fellow AI image makers to dismiss a class-action copyright lawsuit filed by artists. The case is now heading into discovery, meaning plaintiffs' attorneys could soon inspect internal documents and shed light on how the model's datasets were gathered.

Why it matters

We think this editor represents a vital maturation milestone for Midjourney. For designers and content creators, the friction of generating in Discord has always kept the platform feeling like an experimental playground rather than a professional workstation. Bringing a brush-based editor directly to the browser makes fine-tuning images fast enough for commercial turnarounds, and existing users should jump on it immediately.

Strategically, Midjourney is feeling the squeeze. With Flux matching its aesthetic quality and offering better text following, Midjourney can no longer rely solely on prompt-to-image magic to retain users. Owning a polished web editing experience creates product stickiness that Discord bots simply cannot provide.

That said, keeping the web editor gated behind a 10-image Discord requirement shows the team is still hesitant to open the floodgates completely. Combined with an encroaching legal battle over its training data, Midjourney is fighting to modernize its user experience before competitors and regulators close in.