Midjourney API
Midjourney is renowned for its distinctive aesthetics and stylized output. Roko API opens its complete workflow through one unified endpoint — text-to-image, image-to-image, blend, describe, inpaint, variations, and multi-level upscales — billed on a three-dimension grid of task type × speed × quality from $0.011 per task.
- Public model ID
midjourney- Task types
12 (text-to-image through upscale4)- Billing
task type × speed × quality- Endpoint
POST /v1/images/tasks- Catalog snapshot
What is the Midjourney API?
Midjourney is known for stylized finished images, atmospheric lighting, and visual polish. Official V8.2 has been the default since July 24, 2026.
Roko exposes 11 tasks: text-to-image, image-to-image, blend, describe, seed, inpaint, variation1, upsample1, reroll, upscale2, and upscale4. The catalog does not disclose the routed upstream version, so V8 details are official product context only.
Tasks are priced across type, fast/turbo, and standard/hd: 44 rates from $0.011 to $0.594 per async task.
Core capabilities
Distinctive aesthetics
A fit for posters, covers, and concept art.
11 task types
Covers generation, blend, describe, inpaint, variation, and upscale.
Speed and quality controls
fast/turbo and standard/hd balance cost and output.
44 published rates
Prices range from $0.011 to $0.594.
Midjourney examples
Examples below are same-brief comparison assets from a third-party review (PicLumen, August 2026 tests, sources labeled, not Roko-measured), illustrating the model's aesthetic direction; Roko-measured cases will follow once a test key is available.
Illustrative narrative: Odysseus and the olive tree
Odysseus seated beneath an olive tree on a seaside cliff, classical illustration, mythological atmosphere
PicLumen same-brief sample: the fusion of mythological narrative, figure pose, and scene atmosphere shows Midjourney's signature aesthetic treatment.
Third-party same-brief review (PicLumen, Aug 2026) · 2026-10-07 核验
Documentary photo: woman with umbrella in the rain
A woman holding an umbrella on a rain-soaked street at night, cinematic portrait, neon reflections
Rain-night neon, wet-street reflections, and mood handling — the representative direction of Midjourney's cinematic portraiture.
Third-party same-brief review (PicLumen, Aug 2026) · 2026-10-07 核验
Experimental typography: Field Notes poster
Experimental Field Notes design poster with geometric typography, Berlin event layout
The experimental design direction from the review: bold geometric composition and typography showing Midjourney's stylized imagination.
Third-party same-brief review (PicLumen, Aug 2026) · 2026-10-07 核验
Product photography: NOMA perfume
NOMA No. 03 perfume product shot with a glass bottle, sliced fruit and ice, studio lighting
Still-life composition and lighting — glass textures and prop atmosphere sample Midjourney's commercial photography direction.
Third-party same-brief review (PicLumen, Aug 2026) · 2026-10-07 核验
Material texture: felted fox miniature
A felted wool fox with a small bird perched on its head, soft clouds, miniature craft aesthetic
From an AI Primer third-party V8 test article: the fine felted texture and miniature-scene mood sample the V8 family's material rendering.
AI Primer third-party V8 test article screenshot · 2026-10-07 核验
Display typography: GO GET 'EM TIGER
Bold poster design with tiger illustrations and the headline GO GET 'EM TIGER in display type, orange and blue palette
The official showcase of V8's quoted-text rendering: the display headline renders complete and legible alongside tiger artwork and layout hierarchy.
AI Primer third-party V8 test article screenshot · 2026-10-07 核验Midjourney API pricing
Pricing is generated from the current Roko catalog on the task type × speed (fast/turbo) × quality (standard/hd) grid — all 44 tiers listed in the table below. Confirm the console estimate before submitting.
| Resolution | Price | Billing unit |
|---|---|---|
| generation_type=describe · quality=standard · speed=fast | $0.011 | per output |
| generation_type=seed · quality=standard · speed=fast | $0.011 | per output |
| generation_type=describe · quality=standard · speed=turbo | $0.011 | per output |
| generation_type=seed · quality=standard · speed=turbo | $0.011 | per output |
| generation_type=describe · quality=hd · speed=fast | $0.0165 | per output |
| generation_type=seed · quality=hd · speed=fast | $0.0165 | per output |
| generation_type=describe · quality=hd · speed=turbo | $0.0165 | per output |
| generation_type=seed · quality=hd · speed=turbo | $0.0165 | per output |
| generation_type=text-to-image · quality=standard · speed=fast | $0.055 | per output |
| generation_type=image-to-image · quality=standard · speed=fast | $0.055 | per output |
| generation_type=blend · quality=standard · speed=fast | $0.055 | per output |
| generation_type=inpaint · quality=standard · speed=fast | $0.055 | per output |
| generation_type=variation1 · quality=standard · speed=fast | $0.055 | per output |
| generation_type=upsample1 · quality=standard · speed=fast | $0.055 | per output |
| generation_type=reroll · quality=standard · speed=fast | $0.055 | per output |
| generation_type=blend · quality=standard · speed=turbo | $0.055 | per output |
| generation_type=text-to-image · quality=standard · speed=turbo | $0.077 | per output |
| generation_type=image-to-image · quality=standard · speed=turbo | $0.077 | per output |
| generation_type=inpaint · quality=standard · speed=turbo | $0.077 | per output |
| generation_type=variation1 · quality=standard · speed=turbo | $0.077 | per output |
| generation_type=upsample1 · quality=standard · speed=turbo | $0.077 | per output |
| generation_type=reroll · quality=standard · speed=turbo | $0.077 | per output |
| generation_type=text-to-image · quality=hd · speed=fast | $0.0825 | per output |
| generation_type=image-to-image · quality=hd · speed=fast | $0.0825 | per output |
| generation_type=blend · quality=hd · speed=fast | $0.0825 | per output |
| generation_type=inpaint · quality=hd · speed=fast | $0.0825 | per output |
| generation_type=variation1 · quality=hd · speed=fast | $0.0825 | per output |
| generation_type=upsample1 · quality=hd · speed=fast | $0.0825 | per output |
| generation_type=reroll · quality=hd · speed=fast | $0.0825 | per output |
| generation_type=blend · quality=hd · speed=turbo | $0.0825 | per output |
| generation_type=text-to-image · quality=hd · speed=turbo | $0.1155 | per output |
| generation_type=image-to-image · quality=hd · speed=turbo | $0.1155 | per output |
| generation_type=inpaint · quality=hd · speed=turbo | $0.1155 | per output |
| generation_type=variation1 · quality=hd · speed=turbo | $0.1155 | per output |
| generation_type=upsample1 · quality=hd · speed=turbo | $0.1155 | per output |
| generation_type=reroll · quality=hd · speed=turbo | $0.1155 | per output |
| generation_type=upscale2 · quality=standard · speed=fast | $0.132 | per output |
| generation_type=upscale2 · quality=standard · speed=turbo | $0.132 | per output |
| generation_type=upscale2 · quality=hd · speed=fast | $0.198 | per output |
| generation_type=upscale2 · quality=hd · speed=turbo | $0.198 | per output |
| generation_type=upscale4 · quality=standard · speed=fast | $0.396 | per output |
| generation_type=upscale4 · quality=standard · speed=turbo | $0.396 | per output |
| generation_type=upscale4 · quality=hd · speed=fast | $0.594 | per output |
| generation_type=upscale4 · quality=hd · speed=turbo | $0.594 | per output |
Cost examples
The everyday generation tier; image-to-image and blend share the price.
The highest tier is the 4x upscale at hd quality; describe and seed cost just $0.011.
Parameters and supported scope
Fields come from the current generation_config. generation_type selects the task and billing tier; required inputs follow request validation.
| Parameter | Values | Notes |
|---|---|---|
prompt | textarea | Describe the image to create or the changes to apply to references. · Optional |
generation_type | text-to-image / image-to-image / blend / describe / seed / inpaint / variation1 / upsample1 / reroll / upscale2 / upscale4 | Midjourney task type (text-to-image, image-to-image, blend, describe, inpaint, variations, upscales); it selects the billing tier. · Default: text-to-image · Optional |
images | ≤5 | Reference images may use URL or Base64; the current model field shows the applicable limit. · Optional |
mode | fast / turbo | Generation speed tier; faster tiers cost more per request. · Default: fast · Optional |
quality | standard / hd | The quality tier follows the model config (auto / low / medium / standard / hd) and can directly affect price. · Default: standard · Optional |
Call Midjourney with Roko API
Run the async task flow: submit a task to get a task_id, then poll for the result. Never expose an API key in client-side code.
Submit generation
curl https://api.rokoapi.com/v1/images/tasks \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: YOUR_UNIQUE_REQUEST_ID" \
-d '{
"model": "midjourney",
"prompt": "mythological illustration of Odysseus resting beneath an olive tree on a seaside cliff, painterly light",
"quality": "standard",
"mode": "fast",
"generation_type": "text-to-image"
}'Query recovery task
curl https://api.rokoapi.com/v1/tasks/$TASK_ID \
-H "Authorization: Bearer $API_KEY"POST /v1/images/tasks returns 202 with a task_id; once generation completes, query GET /v1/tasks/{task_id} for the result or failure reason. Idempotency-Key must be unique; replays with the same key return the original task without re-billing.
Best-fit scenarios
Posters and covers
Suited to visually led finished work.
Mythic coastal illustration with painterly light.
Blend exploration
blend combines references; describe reads an image.
Blend two mood references.
Final upscaling
Upscale operations enlarge a selected draft.
Upscale the draft for print.
Local revisions
inpaint changes a selected region.
Replace only the background with a dusk sky.
Important limitations
- Task type drives both billing and required inputs: blend needs multiple images, describe/seed are low-cost query tasks, and upscale tiers require a prior generation.
- The 44 tiers vary with task type, speed, and quality combinations — price batch runs by target tier first.
- The upstream version identifier is not disclosed in the Roko catalog; the live upstream follows the task response, and this page does not equate it with any specific official version.
- A submitted task is not a successful generation — trust the task query result, and review outputs before use.
Midjourney or another image model?
Choose Midjourney for stylization, Grok Image 2.0 for typography and editing, and GPT Image 2 for realism. Roko does not disclose the upstream Midjourney version; follow the catalog and actual result.
All models →Midjourney API FAQ
What model ID should I use for the Midjourney API?
Use the public Roko API model ID midjourney through the unified endpoint — no Midjourney subscription or Discord required.
How much does the Midjourney API cost?
Billing runs on task type × speed × quality across 44 tiers: describe/seed from $0.011 per task, everyday text-to-image (standard/fast) at $0.055, and upscale4 (hd) topping out at $0.594. Prices can change; confirm the estimate and final settlement.
What are the 11 task types?
text-to-image, image-to-image, blend (multi-image fusion), describe (image description), seed, inpaint (local repaint), variation1, upsample1, reroll, and the upscale2/upscale4 escalation levels.
How do I choose between fast/turbo and standard/hd?
fast is standard speed and turbo is faster but pricier; standard is the base quality tier and hd is higher but costs more. For the same task type, turbo and hd tiers usually price above the fast/standard pair — the full grid is in the pricing table.
Which aspect ratios are supported?
Midjourney controls framing through prompt parameters rather than a dedicated form field; the Roko form currently exposes no aspect-ratio field, so framing follows request validation.
Do I need a Midjourney subscription?
No. Roko API bills per task through the unified endpoint — no official Midjourney subscription or Discord account needed.
Which endpoint does it use?
Submit async tasks via POST /v1/images/tasks to get a task_id, then poll GET /v1/tasks/{task_id} for the result or failure reason.
Which upstream Midjourney version does Roko use?
The Roko catalog does not disclose an upstream version identifier — the live upstream follows the task response. On the official side, the V8 series has evolved since the March 2026 Alpha, and V8.2 has been the official default since July 24, 2026.