Midjourney

Midjourney is an AI image and video generation tool, created by an independent research lab of the same name founded in 2021 by David Holz in San Francisco. Unlike most AI labs, Midjourney became profitable without outside venture investors within its first year and has stayed largely self-funded since, reportedly earning around $300 million in revenue by 2024.

Midjourney generates images and short videos from text prompts, reference images, or a combination of both, and has moved beyond its original Discord-only interface to a full web app at midjourney.com, which adds an image editor (inpainting, cropping, resizing, outpainting) not available on Discord. The current default model is V8.2 (released July 24, 2026), following V8.1 (April 2026), which introduced roughly 4-5x faster standard rendering, HD 2K output without upscaling, and improved prompt adherence; V9 is in early training as a scaled-up continuation of V8 rather than a fundamentally different architecture. Video generation animates a generated or uploaded image into a clip starting at 5 seconds and extendable up to 21 seconds. Midjourney also runs Patchwork, a collaborative, LLM-powered infinite canvas for up to 100 users to co-build characters, scenes, and stories together in real time.

In June 2026 Midjourney announced a striking pivot beyond image generation: Midjourney Medical, a division building a full-body ultrasound scanner ("Ultrasonic CT") that images the whole body in about 60 seconds using roughly 500,000 ultrasonic sensors, without radiation or MRI-style magnets, with a flagship "Midjourney Spa" scanning location planned for San Francisco by the end of 2027. This entry covers Midjourney's core AI image and video generation product, not the separate Medical division.

1. Sign up at midjourney.com (a Google account works) — there's no free trial, so a paid plan (starting at Basic, $10/month) is required from the start.
2. Generate images or short videos from a text prompt, an uploaded reference image, or both, using the web app or the original /imagine command in Midjourney's Discord server.
3. Use the web editor for inpainting, outpainting, cropping, resizing, and changing aspect ratios on a generated image — capabilities only available on the web, not on Discord.
4. Animate a generated or uploaded image into a short video clip, starting at 5 seconds and extendable up to 21 seconds.
5. Manage your monthly Fast GPU hours: use Fast mode for quick turnaround on priority jobs, or switch to unlimited Relax mode (Standard plan and above) for slower, queued generation once your Fast hours run out.
6. For collaborative worldbuilding, open Patchwork's infinite canvas to add characters, locations, events, and props together with up to 99 other people in real time.

• Generating images from a text prompt, a reference image, or both — for concept art, illustration, marketing visuals, or personal creative projects.
• Editing a generated image directly in the web app — inpainting, outpainting, cropping, or resizing — without exporting to a separate tool.
• Animating a still image into a short video clip (5-21 seconds) for social content, mood boards, or previsualization.
• Collaborative worldbuilding on Patchwork's shared infinite canvas, with up to 100 people building out characters, locations, and story elements together in real time.
• High-volume creative production on the Pro or Mega plans, which include Stealth Mode to keep generations out of Midjourney's public gallery.
• Studios and teams that want a self-funded, non-VC-backed AI lab as their image/video vendor, given Midjourney's unusual funding history.