Midjourney

Midjourney — Aesthetic-First AI Image & Video Studio (V7 Default, Web App Primary, Active, Not Shut Down)

Not shut down.​ Midjourney (midjourney.com, San Francisco independent lab, David Holz, founded 2021/2022, ~10M+ active users, 500M+ gens/day) active 2026-08, no discontinuation. V7 is current default (Apr 2026), web app at midjourney.com is now primary surface (Discord bot still works, no longer mandatory), Niji 7 (Jan 2026) for anime/illustration, V8.1 alpha on alpha.midjourney.com only. Category ai-productivity-data; pipeline_stage none. Killer differentiator = unmatched artistic/editorial aesthetic + V7 character/style consistency (--cref/--sref/Omni Reference) + Draft Mode (10× faster, ½ GPU) + Relax unlimited on Standard+ + moodboards/personalization, not Leonardo (asset-bench + custom training + canvas, weaker hero aesthetics), not DALL·E 3/GPT Image (literal prompt adherence + text-in-image, less art-directed), not Stable Diffusion/Flux (open-weight/local/API-first, no hosted aesthetic tuning), not Adobe Firefly (IP-safe but conservative look). 2026 trap: no free tier ever since 2023​ (floor $10 Basic = 3.3 fast GPU-hr ≈ 200 imgs, no Relax, once fast burns you wait 30 days), no official public API​ (ToS bans automation, third-party wrappers = ban risk), Stealth Mode private gens locked to $60 Pro+​ (Basic/Standard gens public in community gallery by default), text-in-image still unreliable vs Ideogram/Recraft (V7 ~75% English / ~50% CJK), no custom LoRA training, no inpainting as deep as Leonardo/Photoshop, fast GPU hrs don’t roll over, $4/hr top-up.

1. One-line positioning

Midjourney = the beauty engine, not the production bench: log into midjourney.com (or /imagine in Discord) → type a rich prompt → V7 returns 4 art-directed variants (8–10s std, Draft Mode ~1s ½ cost) → --cref locks a character across scenes, --sref borrows a style, Omni Reference drags subject/style/structure layers, Personalization learns your rated taste, Moodboards steer direction → upscale (1x/2x), Vary Region (inpainting), Pan/Zoom, Remix → on Pro/Mega animate stills to SD/HD video (Relax video on Pro+) → Standard+ gets unlimited Relax queue (slower, no fast burn), Pro+ gets Stealth (private) + 12 fast concurrency. Fits ai-productivity-data; pipeline_stage none (prompt→variants→refine→upscale→animate). For concept art, editorial, brand mood, album/book cover, game key art — it’s still the aesthetic benchmark; for on-brand asset batches with consistent mascot + SVG handoff, Leonardo/Recraft win; for signage with legible text, Ideogram.

2. Core features

3. Pricing (2026-08, USD — GPU-time metered, annual 20% off upfront)

pricing_type = subscription-gpu-time-metered-no-free-tier; starting_price = $10; free_tier = "none; trial removed 2023, every generation requires paid sub, floor $10 Basic 3.3 fast GPU-hr ~200 imgs no relax no stealth"; pricing_note = "midjourney-sf-holz-2021-2022-active-2026-no-shutdown; v7-default-apr-2026-web-primary-discord-optional-not-shutdown; basic-10-3.3hr-no-relax-fast-only-burns-out-in-afternoon; standard-30-sweet-spot-freelance-relax-unlimited; pro-60-stealth-mandatory-client-nd-a-company-over-1m-rev; mega-120-studios; fast-hr-no-rollover; extra-fast-4-per-hr; draft-mode-10x-0.5x-gpu; video-sd-8x-image-gpu-hd-26x; no-official-api-tos-bans-automation-third-party-wrappers-ban-risk"

4. Access Type

access_type: web-app-plus-discord-bot-no-native-mobile-no-desktop-no-public-api

access_display: midjourney.com (Create/Organize/Edit/Explore, V8.1 alpha on alpha.midjourney.com) · 💬 Discord /imagine DM or server (still works, lags web on newest params) · no iOS/Android native app (responsive web only), no Electron desktop, no official REST API​ (ToS prohibits automated access; all “MJ APIs” are unofficial wrappers)

5. Reviews (2026)

Praise: aesthetic coherence unmatched​ (moody light, texture, composition intentional not generic), V7 fixed hands/long-prompt/char-consistency, web studio (gallery/pan/zoom/vary-region) removed Discord tax, Draft Mode 10×/½ cost makes breadth-cheap, --cref/--sref/Omni Ref give reusable direction, Relax unlimited on $30, Niji 7 best anime model, community Explore is a prompt-university.

Gripes: no free tier​ (ChatGPT/DALL·E/SD/Flux all have one), $10 Basic no Relax = 200 imgs then stall, Stealth private locked $60+, no official API​ (devs must wrap Discord = ToS risk), text-in-image ~75% English / ~50% CJK still worse than Ideogram/Recraft, no custom LoRA training/export, fast GPU hrs don’t roll over, prompt adherence aesthetic-not-literal (product mockups drift), video is silent/clip-only/no lip-sync, no vector/SVG, commercial >$1M rev forced to Pro+.

6. Best for / Not for

7. Competitors