Does Midjourney V7 have an official API?

Midjourney does not publish an official public API. reApi exposes Midjourney V7 as a managed async API: one key, native MJ prompt syntax, a task id per request, and per-request billing. You POST a prompt and poll the task for the four results.

How is Midjourney V7 billed?

Per request, not per image — each Midjourney V7 request produces four images and is charged once. Price scales with the speed tier (draft / fast / turbo) and the operation (generation vs lightweight vs heavy edit). Using Omni Reference (--oref) doubles the cost. See the pricing table above for current rates.

How does Midjourney V7 image-to-image work?

Place one or more public image URLs at the start of the prompt, followed by your text. One image plus text, or two-plus images with or without text, are valid; a single image with no text is rejected. Tune reference influence with --iw. reApi accepts only public http(s) image URLs — no base64.

Which Midjourney parameters does V7 accept?

Midjourney V7 accepts --ar (aspect ratio), --s (stylize 0-1000), --chaos (0-100), --no (negative prompt), --seed, --sref (Style Reference), --sw (style weight), --oref (Omni Reference), --iw (image weight), and --sv (style version). The version is locked to V7, so --v and --niji are ignored; speed flags go in the speed field, not the prompt.

What edit operations can I run on a Midjourney V7 image?

Eleven: upscale, variation, remix, enhance, pan, outpaint, inpaint, canvas edit, background removal, retexture, and upload-based repaint. The first eight reference a completed Midjourney V7 task by id; the last three take a fresh public image URL. Each is a separate model id on the same endpoint.

How do I upscale or vary a specific Midjourney V7 image?

Pass the reApi task id of your generation plus image_number (0-3) to pick which of the four grid images to act on. For example, model mj-v7-upscale with model_params carrying task_id and image_number 0 upscales the first image. reApi translates your task id to the upstream reference automatically.

Is Midjourney V7 good for product photography?

Yes. Midjourney V7 handles hero shots, ecommerce packshots, mockups, and ad creatives. Pair a clean prompt with Omni Reference for subject anchoring and Style Reference for consistent lighting and material direction — the standard Midjourney V7 setup for brand-consistent commercial work.

Is Midjourney V7 free or paid?

Paid. Completed generations are charged at the rates in this page's pricing table. Eligible new accounts may receive promotional test credits; availability and amount are shown at signup or in the dashboard. Technical or provider failures are refunded automatically, except where this page documents a charged output-moderation block.

Does Midjourney V7 require a monthly subscription?

No. Pricing is pay-as-you-go: $1 buys 1,000 credits and credits never expire. Cost scales with what you generate rather than with a recurring fee.

How do I get a Midjourney V7 API key?

Sign up and create a key in the dashboard, no card required. Pass it as a Bearer token in the Authorization header.

Midjourney V7 — Image Generation, Upscale & Editing

Midjourney V7 turns a text prompt into four high-quality images per request, with native --ar / --s / --chaos syntax, Omni Reference (--oref), and Style Reference (--sref). Call Midjourney V7 from one OpenAI-style endpoint on reApi — plus eleven edit operations: upscale, variation, remix, inpaint, outpaint, pan, enhance, retexture, canvas edit, and background removal.

Highlights: 12 operations · draft / fast / turbo · pan, zoom, vary, edit
Best for: Concept art, stylized illustration
Input: Text, reference image or prior task
Output: Images (grid or single, by operation)

Input

Prompt

Describe the image; append MJ flags like --ar 16:9 --s 500

Speed

Affects price and queue priority

Estimated cost~$0.08888 credits

Result

Try one of these prompts

What is the Midjourney V7 API?

Midjourney V7 is Midjourney's image model, exposed on reAPI as 12 operations on one async endpoint: generation plus 11 edits — variation, upscale, remix, enhance, pan, outpaint, inpaint, canvas edit, remove background, retexture and mask edit. Native Midjourney flags such as --ar, --s, --chaos and --sref stay inside the prompt, every generation returns 4 images at draft, fast or turbo speed, and billing is per request.

Midjourney V7 image generation grid of four variations

What you can build with this model

Real-world workflows and production use cases you can build and ship with this model.

Four Midjourney V7 image variations from one prompt

Generate four Midjourney V7 variations per request

Every Midjourney V7 request returns up to four image variations, so a single call gives an art director real options instead of one take. Drive composition with native flags — --ar for aspect ratio, --s for stylize, --chaos for spread — and pick the frame that lands. Because Midjourney V7 is async, you submit a prompt, get a task id back, and poll for the four results when they are ready.

Open the playground

Midjourney V7 Omni Reference keeping a product consistent across shots

Anchor a subject with Omni Reference and Style Reference

Midjourney V7 supports Omni Reference (--oref) to keep a character, product, or object consistent across generations, and Style Reference (--sref) with an adjustable --sw weight to lock a look. That is the right control set for brand-consistent product photography and reusable style systems — the same Midjourney V7 reference setup repeated across an entire campaign.

Midjourney V7 upscale and inpaint edit operations on a generated image

Edit without leaving the Midjourney V7 pipeline

Once Midjourney V7 generates an image, the same task id flows into eleven edit operations: upscale to high resolution, pull subtle or strong variations, remix with a new prompt, inpaint a masked region, outpaint or pan the canvas, enhance a draft, retexture, or strip the background. No re-upload, no second model — the edit operations reference the Midjourney V7 task you already ran.

Pricing

Credit-based — 1 credit = $0.001 USD. Pay only for completed generations.

First test cost

$0.04444 credits

One sample at the cheapest tier (1 request (4 images)).

Testing budget guide

Add credits

$10

≈ 229 tests

$50

≈ 1145 tests

$100

≈ 2290 tests

Model	Category	Price
Draft	Image generation	$0.044 44 credits · 1 request (4 images)
Fast	Image generation	$0.088 88 credits · 1 request (4 images)
	Lightweight edit (remove-bg / pan / outpaint / enhance)	$0.088 88 credits · 1 request
	Heavy edit (upscale / variation / remix / inpaint / canvas)	$0.131 131 credits · 1 request
Turbo	Image generation	$0.175 175 credits · 1 request (4 images)
	Lightweight edit (remove-bg / pan / outpaint / enhance)	$0.175 175 credits · 1 request
	Heavy edit (upscale / variation / remix / inpaint / canvas)	$0.262 262 credits · 1 request

Why reAPI

Four images, one async call

Midjourney V7 bills per request, not per image — each call produces four variations. Submit a prompt to /api/v1/images/generations, receive a task id, and poll until the four images are ready. The API returns 1-4 images depending on content review, and you are charged once per request regardless.

Native Midjourney syntax, verbatim

Use Midjourney V7 parameters directly in the prompt — --ar, --s, --chaos, --no, --seed, --sref, --oref, --iw, --sw, --sv. reApi forwards them unchanged, so prompts you already wrote for Midjourney V7 work as-is. The version is locked to V7; --v and --niji are ignored.

Generation and editing in one surface

Midjourney V7 on reApi is not just text-to-image. The same API key reaches upscale, variation, remix, inpaint, outpaint, pan, enhance, retexture, canvas edit, and background removal — eleven Midjourney V7 edit operations that chain off any image you generate.

Midjourney V7 vs DALL·E 3

Midjourney V7 and DALL·E 3 are both premium text-to-image models. Here is how Midjourney V7 on reApi is positioned against DALL·E 3 on the dimensions that matter for production image pipelines.

Capability

Midjourney V7 on reApi

DALL·E 3

Images per request

Four variations per request — pick the best frame from one call.

One image per request; multiple takes mean multiple calls.

Reference control

Omni Reference (--oref) for subject consistency and Style Reference (--sref / --sw) for look transfer.

No first-class subject/style reference; control is prompt-only.

Native syntax

Full MJ flag set in the prompt — --ar, --s, --chaos, --seed, --iw, and more.

Plain natural-language prompts; no parameter syntax.

Edit operations

Upscale, variation, remix, inpaint, outpaint, pan, enhance, retexture, background removal — eleven in all.

Inpainting via a separate edit endpoint; no upscale/variation chain.

Billing model

Per request (four images), with a draft / fast / turbo speed-cost tradeoff.

Per image, by resolution and quality.

Integration

Async task id + poll on an OpenAI-style endpoint; one key for generation and all edits.

Synchronous image response on the OpenAI Images API.

Comparison reflects publicly documented behavior at the time of writing. Capabilities and pricing on both sides change; verify current details against each provider before integrating.

Call Midjourney V7 in three steps

step 01
Create an API key
Sign up at reApi and grab a key. Free signup credits cover dozens of Midjourney V7 generations — enough to test image-to-image, Omni Reference, and an upscale before topping up.
Open
step 02
Submit a Midjourney V7 prompt
POST a request with model mj-v7 and your prompt to /api/v1/images/generations, with native MJ flags in the prompt. Midjourney V7 returns a task id immediately — generation is asynchronous and typically completes in one to three minutes.
Open
step 03
Poll for the four results
GET /api/v1/tasks/:task_id until status is completed, then read the image URLs. Feed that same task id into any Midjourney V7 edit operation — upscale, variation, inpaint — to keep working on the result.
Open

docs/api/midjourney-v7

API reference

Drop-in code and the full parameter table.

View full reference

curl -X POST https://reapi.ai/api/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "mj-v7",
    "prompt": "a serene Japanese garden with cherry blossoms --ar 16:9 --s 500",
    "model_params": { "speed": "fast" }
  }'

Parameter validation failed

Rejected synchronously with the failing field named. Check enum values, ranges, and URL fields — nothing is charged.

Authentication or balance issue

401 means a missing or invalid Bearer key; 402 means the reserve exceeds your balance. Manage keys and credits in the console.

Content or material rejected

Reference material and outputs pass automated moderation. Rejected tasks fail with a clear error and a full refund.

Task failed or timed out

A task that reaches a failed state is never charged — the reserve refunds automatically. Keep the task id and retry when ready.

Frequently asked questions

Common questions about this model.

Midjourney V7 is Midjourney’s V7 image generation model. It renders four images per request from native MJ prompt syntax, supports Omni Reference and Style Reference, and chains into eleven edit operations. On reApi it ships as model id mj-v7 on the unified /api/v1/images/generations endpoint, with the edits as sibling model ids like mj-v7-upscale and mj-v7-inpaint.