Midjourney

What Midjourney does

Midjourney turns written prompts into still images, and more recently into short animated clips built from those images. It began as a Discord bot and now runs mainly through a web application with a prompt bar, a personal image feed, and an editor for reframing, retexturing and extending existing pictures. Its output leans toward a consistent house aesthetic: painterly light, shallow depth of field, and strong composition even from short prompts. Parameters control aspect ratio, stylisation strength, variety and model version, and style references let you carry a look across a series. It suits concept artists, illustrators, art directors and anyone who wants images that look deliberate rather than generic. The main limitation is control: getting an exact layout, a specific product, or accurate text inside the image takes far more iteration than it does in tools designed around editing and prompt fidelity.

How to use it

You sign up at midjourney.com with a Discord or Google account and pick a plan, since there is no permanent free tier. The first run is a prompt typed into the create bar; four images come back in under a minute, and you pick one to upscale, vary, or open in the editor. Most work happens in cycles: write a short prompt describing subject, setting and lighting, then adjust with parameters such as aspect ratio and stylisation rather than piling on adjectives. Style references and character references are where the tool becomes genuinely useful for series work, so learn those early. Two things commonly trip people up. First, your images are public by default on the shared feed unless your plan includes stealth mode, which matters for client work. Second, plans buy fast GPU hours rather than a fixed image count, and heavy iteration drains them quickly; relaxed mode queues generations more slowly but does not consume fast hours.

Best for

• concept art
• illustration
• mood boards
• stylised visuals
• series with a consistent look

Strengths and limitations