Midjourney

POST

tasks

Midjourney

curl --request POST \
  --url https://api.qingbo.ai/v1/tasks \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "action": "<string>",
  "prompt": "<string>",
  "aspect_ratio": "<string>",
  "image_urls": [
    "<string>"
  ],
  "seed": 123,
  "negative_prompt": "<string>",
  "speed": "<string>",
  "version": "<string>",
  "ref_task_id": "<string>",
  "index": 123,
  "custom_id": "<string>",
  "zoom_ratio": 123,
  "direction": "<string>",
  "mask_url": "<string>",
  "metadata": {},
  "style": "<string>",
  "stylize": 123,
  "chaos": 123,
  "weird": 123,
  "tile": true,
  "iw": 123,
  "cref": "<string>",
  "cw": 123,
  "sref": "<string>",
  "sw": 123,
  "dref": "<string>",
  "dw": 123,
  "raw": true,
  "draft": true,
  "hd": true,
  "stop": 123,
  "extra": "<string>"
}
'

import requests

url = "https://api.qingbo.ai/v1/tasks"

payload = {
    "model": "<string>",
    "action": "<string>",
    "prompt": "<string>",
    "aspect_ratio": "<string>",
    "image_urls": ["<string>"],
    "seed": 123,
    "negative_prompt": "<string>",
    "speed": "<string>",
    "version": "<string>",
    "ref_task_id": "<string>",
    "index": 123,
    "custom_id": "<string>",
    "zoom_ratio": 123,
    "direction": "<string>",
    "mask_url": "<string>",
    "metadata": {},
    "style": "<string>",
    "stylize": 123,
    "chaos": 123,
    "weird": 123,
    "tile": True,
    "iw": 123,
    "cref": "<string>",
    "cw": 123,
    "sref": "<string>",
    "sw": 123,
    "dref": "<string>",
    "dw": 123,
    "raw": True,
    "draft": True,
    "hd": True,
    "stop": 123,
    "extra": "<string>"
}
headers = {"Content-Type": "application/json"}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {'Content-Type': 'application/json'},
  body: JSON.stringify({
    model: '<string>',
    action: '<string>',
    prompt: '<string>',
    aspect_ratio: '<string>',
    image_urls: ['<string>'],
    seed: 123,
    negative_prompt: '<string>',
    speed: '<string>',
    version: '<string>',
    ref_task_id: '<string>',
    index: 123,
    custom_id: '<string>',
    zoom_ratio: 123,
    direction: '<string>',
    mask_url: '<string>',
    metadata: {},
    style: '<string>',
    stylize: 123,
    chaos: 123,
    weird: 123,
    tile: true,
    iw: 123,
    cref: '<string>',
    cw: 123,
    sref: '<string>',
    sw: 123,
    dref: '<string>',
    dw: 123,
    raw: true,
    draft: true,
    hd: true,
    stop: 123,
    extra: '<string>'
  })
};

fetch('https://api.qingbo.ai/v1/tasks', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.qingbo.ai/v1/tasks"

	payload := strings.NewReader("{\n  \"model\": \"<string>\",\n  \"action\": \"<string>\",\n  \"prompt\": \"<string>\",\n  \"aspect_ratio\": \"<string>\",\n  \"image_urls\": [\n    \"<string>\"\n  ],\n  \"seed\": 123,\n  \"negative_prompt\": \"<string>\",\n  \"speed\": \"<string>\",\n  \"version\": \"<string>\",\n  \"ref_task_id\": \"<string>\",\n  \"index\": 123,\n  \"custom_id\": \"<string>\",\n  \"zoom_ratio\": 123,\n  \"direction\": \"<string>\",\n  \"mask_url\": \"<string>\",\n  \"metadata\": {},\n  \"style\": \"<string>\",\n  \"stylize\": 123,\n  \"chaos\": 123,\n  \"weird\": 123,\n  \"tile\": true,\n  \"iw\": 123,\n  \"cref\": \"<string>\",\n  \"cw\": 123,\n  \"sref\": \"<string>\",\n  \"sw\": 123,\n  \"dref\": \"<string>\",\n  \"dw\": 123,\n  \"raw\": true,\n  \"draft\": true,\n  \"hd\": true,\n  \"stop\": 123,\n  \"extra\": \"<string>\"\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.qingbo.ai/v1/tasks")
  .header("Content-Type", "application/json")
  .body("{\n  \"model\": \"<string>\",\n  \"action\": \"<string>\",\n  \"prompt\": \"<string>\",\n  \"aspect_ratio\": \"<string>\",\n  \"image_urls\": [\n    \"<string>\"\n  ],\n  \"seed\": 123,\n  \"negative_prompt\": \"<string>\",\n  \"speed\": \"<string>\",\n  \"version\": \"<string>\",\n  \"ref_task_id\": \"<string>\",\n  \"index\": 123,\n  \"custom_id\": \"<string>\",\n  \"zoom_ratio\": 123,\n  \"direction\": \"<string>\",\n  \"mask_url\": \"<string>\",\n  \"metadata\": {},\n  \"style\": \"<string>\",\n  \"stylize\": 123,\n  \"chaos\": 123,\n  \"weird\": 123,\n  \"tile\": true,\n  \"iw\": 123,\n  \"cref\": \"<string>\",\n  \"cw\": 123,\n  \"sref\": \"<string>\",\n  \"sw\": 123,\n  \"dref\": \"<string>\",\n  \"dw\": 123,\n  \"raw\": true,\n  \"draft\": true,\n  \"hd\": true,\n  \"stop\": 123,\n  \"extra\": \"<string>\"\n}")
  .asString();

Midjourney is known for highly aesthetic, artistic output, renders 2K images by default, and covers anime styles through its niji versions. There is a single model ID, midjourney, and action decides what a call does: imagine turns a prompt into a grid of four images, and every other action continues from the result of a previous task. All actions are submitted through the unified POST /v1/tasks endpoint.

Quick start

Action	What it does	Required parameters	Images
`imagine`	Turns a prompt into a grid of four; add `image_urls` for image-guided generation	`prompt`	4
`upscale`	Takes one image out of the grid as a single image	`ref_task_id`, `index`	1
`variation`	Small variation of one image in the grid	`ref_task_id`, `index`	4
`high-variation`	Large variation of a single image; composition and style may both change	`ref_task_id`, `index`	4
`low-variation`	Small variation of a single image, keeping subject and tone	`ref_task_id`, `index`	4
`reroll`	Draws a new grid from the source task’s prompt	`ref_task_id`	4
`zoom`	Zooms out, keeping the original and filling in more background	`ref_task_id`, `index`	4
`pan`	Extends the image in one direction; repeat to build a panorama	`ref_task_id`, `index`, `direction`	4
`inpaint`	First step of inpainting: picks the single image to repaint	`ref_task_id`, `index`	via `modal`
`modal`	Adds the mask and prompt to an `inpaint` task and finishes the repaint	`ref_task_id`, `mask_url`	4
`edit`	Rewrites a whole image from a prompt	`ref_task_id`, `index`, `prompt`, `image_urls`	1–4
`remix-strong`	Strong reshape with a new prompt; `v8.1` / `v8.2` only	`ref_task_id`, `index`	4
`remix-subtle`	Subtle reshape with a new prompt; `v8.1` / `v8.2` only	`ref_task_id`, `index`	4

The number of images is whatever result.images contains once the task completes. Every action except imagine takes a ref_task_id — the task_id of the previous task. See Examples for how the chain fits together.

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "imagine",
    "prompt": "a lighthouse on a cliff at dawn, painterly, no text",
    "aspect_ratio": "16:9",
    "version": "v7",
    "speed": "fast"
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "upscale",
    "ref_task_id": "task-wave1789004688b950049238",
    "index": 2
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "variation",
    "ref_task_id": "task-wave1789004688b950049238",
    "index": 3,
    "speed": "fast"
  }'

A successful submission returns a task_id. Retrieve the result with GET /v1/tasks/{task_id}, wait inline with Prefer: wait, or configure a webhook.

Request parameters

string

required

Must be midjourney.

string

default:"imagine"

What the call does; see the table above for the values.

string

The prompt. Required for imagine and edit; optional for modal, remix-strong and remix-subtle, which fall back to the source task’s prompt when it is empty. Native Midjourney flags can be written into the prompt (for example --ar 16:9); a structured parameter of the same name wins.

string

Output aspect ratio: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9 or 9:21.

string[]

Reference image URLs. On imagine they make the generation image-guided; on edit they are required and hold the image being rewritten.

integer

Random seed. imagine only.

string

Elements to exclude. imagine only.

string

default:"relax"

Generation speed tier: relax, fast or turbo. It directly changes the price and applies to every action.

string

Midjourney model version: v5.1, v5.2, v6.1, v7, v8.1, v8.2, niji6 or niji7. Chained actions inherit the version of their source task.

string

The task_id of the source task, meaning the previous step. Required for every action except imagine.

integer

Which image of the source grid to act on, 1–4 (top-left, top-right, bottom-left, bottom-right). Required for upscale, variation, high-variation, low-variation, zoom, pan, inpaint, edit, remix-strong and remix-subtle.

string

Names the button on the source task directly; it wins over index when both are present. Available on upscale, variation, high-variation, low-variation, reroll, zoom, pan and inpaint.

number

zoom only. Zoom-out factor: below 2 zooms out 1.5×, 2 and above zooms out 2×.

string

pan only. Direction to extend the image: left, right, up or down.

string

modal only. URL of the mask image: transparent areas are repainted, white areas are kept. Required to repaint a region; without it the task is treated as an outpaint.

object

Custom metadata object, passed upstream as given.

string

imagine only. Style switch (--style), for example raw.

integer

imagine only. Stylization strength (--s), 0–1000.

integer

imagine only. How different the four results are from each other (--c), 0–100.

integer

imagine only. Weirdness (--w), 0–3000.

boolean

imagine only. Produces a seamlessly tileable pattern (--tile).

number

imagine only. Image weight (--iw), 0–3. Higher values stay closer to image_urls.

string

imagine only. Character reference image URL (--cref).

integer

imagine only. How closely the character reference is followed (--cw), 0–100. Used with cref.

string

imagine only. Style reference image URL (--sref).

integer

imagine only. How closely the style reference is followed (--sw), 0–1000. Used with sref.

string

imagine only. Depth reference image URL (--dref).

number

imagine only. How closely the depth reference is followed (--dw), 0–100. Used with dref.

boolean

imagine only. Raw style (--raw), available from v5.1.

boolean

imagine only. Draft mode (--draft), available from v7.

boolean

imagine only. HD output (--hd), v8.1 and v8.2 only.

integer

imagine only. Stop early (--stop), 10–100, on v5.1, v5.2, v6.1 and niji6 only.

string

imagine only. Native Midjourney flags appended to the end of the prompt as given.

See Submit Task for callback_url, callback_events, Prefer: wait, Idempotency-Key, and the maximum-cost header. This model does not support n, resolution or quality. Any other parameter returns 400 and is not billed.

Limits

Condition	Limit
Per task	Midjourney returns one fixed image set per task; n is not supported
Any action other than `imagine`	`ref_task_id` is required
`ref_task_id`	Must be your own Midjourney task in status `completed`, otherwise the request returns `400`
`index`	`1`–`4`
Source task of `zoom`, `pan` and `inpaint`	Must be the single image produced by `upscale`; upscale a grid first
`pan`	Supported on `v6.1`, `v7`, `v8.1`, `v8.2` and `niji6`
`remix-strong` and `remix-subtle`	The source task must be `v8.1` or `v8.2`
`inpaint`	Produces no image on its own; call `modal` with a mask to finish it

Pricing

Billed at a fixed price per call across two dimensions: the action and the speed tier. Every action has its own price, and relax, fast and turbo price differently within one action. An imagine call returns four images and is billed once, not per image. Rates are in price_config on GET /v1/models and in the console’s Model Market. What a single call actually cost is the cost field on the task response — an integer quota at 500,000 quota = 1 USD, not dollars. version, aspect ratio and prompt length have no effect on the price, and reference images add no fee, so the cost of a call is known before it is sent. An action with no price key of its own settles at the fallback price. You are only billed for a successfully generated image. Failed and cancelled tasks, and tasks that return no usable image, are refunded in full.

Response

{
  "task_id": "task-wave1789004688b950049238",
  "model": "midjourney",
  "action": "imagine",
  "status": "completed",
  "progress": "100%",
  "created_at": 1788771370,
  "completed_at": 1788771447,
  "billing_status": "settled",
  "cost": 25335,
  "result": {
    "images": [
      { "expires_at": 1789089508, "url": ["https://cdn.example.com/1.png"] },
      { "expires_at": 1789089508, "url": ["https://cdn.example.com/2.png"] },
      { "expires_at": 1789089508, "url": ["https://cdn.example.com/3.png"] },
      { "expires_at": 1789089508, "url": ["https://cdn.example.com/4.png"] }
    ]
  },
  "urls": {
    "get": "https://api.qingbo.ai/v1/tasks/task-wave1789004688b950049238",
    "cancel": "https://api.qingbo.ai/v1/tasks/task-wave1789004688b950049238/cancel"
  }
}

{
  "task_id": "task-wave1789004901b950221487",
  "model": "midjourney",
  "action": "upscale",
  "status": "completed",
  "progress": "100%",
  "created_at": 1788773131,
  "completed_at": 1788773155,
  "billing_status": "settled",
  "cost": 30960,
  "result": {
    "images": [
      { "expires_at": 1789089508, "url": ["https://cdn.example.com/upscaled.png"] }
    ]
  },
  "urls": {
    "get": "https://api.qingbo.ai/v1/tasks/task-wave1789004901b950221487",
    "cancel": "https://api.qingbo.ai/v1/tasks/task-wave1789004901b950221487/cancel"
  }
}

result.images returns the images in grid order, which is the order index refers to: imagine returns four entries, upscale returns one. url is an array and expires_at is when the image link expires. See Query Task Status for the full field reference.

Examples

Chained actions are linked by ref_task_id, which carries the task_id of your own previous task — not an image URL, and not an upstream ID.

Generate a grid

Call imagine and keep the task_id from the response. This step returns four entries in result.images.

Pick one and continue

Pass that task_id as ref_task_id and choose the image with index. upscale gives a single image; variation gives a new grid of four.

Keep going on the single image

The task ID from upscale becomes the ref_task_id of the next step: zoom to expand, pan to extend sideways, or inpaint plus modal to repaint a region. The chain can continue as long as you need.

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "zoom",
    "ref_task_id": "task-wave1789004901b950221487",
    "index": 1,
    "zoom_ratio": 1.5
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "pan",
    "ref_task_id": "task-wave1789004901b950221487",
    "index": 1,
    "direction": "right"
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "inpaint",
    "ref_task_id": "task-wave1789004901b950221487",
    "index": 1
  }'

curl -X POST https://api.qingbo.ai/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "modal",
    "ref_task_id": "task-wave1789004955b950663104",
    "prompt": "replace the selected area with a red leather sofa",
    "mask_url": "https://cdn.example.com/mask.png"
  }'

Midjourney

curl --request POST \
  --url https://api.qingbo.ai/v1/tasks \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "action": "<string>",
  "prompt": "<string>",
  "aspect_ratio": "<string>",
  "image_urls": [
    "<string>"
  ],
  "seed": 123,
  "negative_prompt": "<string>",
  "speed": "<string>",
  "version": "<string>",
  "ref_task_id": "<string>",
  "index": 123,
  "custom_id": "<string>",
  "zoom_ratio": 123,
  "direction": "<string>",
  "mask_url": "<string>",
  "metadata": {},
  "style": "<string>",
  "stylize": 123,
  "chaos": 123,
  "weird": 123,
  "tile": true,
  "iw": 123,
  "cref": "<string>",
  "cw": 123,
  "sref": "<string>",
  "sw": 123,
  "dref": "<string>",
  "dw": 123,
  "raw": true,
  "draft": true,
  "hd": true,
  "stop": 123,
  "extra": "<string>"
}
'

import requests

url = "https://api.qingbo.ai/v1/tasks"

payload = {
    "model": "<string>",
    "action": "<string>",
    "prompt": "<string>",
    "aspect_ratio": "<string>",
    "image_urls": ["<string>"],
    "seed": 123,
    "negative_prompt": "<string>",
    "speed": "<string>",
    "version": "<string>",
    "ref_task_id": "<string>",
    "index": 123,
    "custom_id": "<string>",
    "zoom_ratio": 123,
    "direction": "<string>",
    "mask_url": "<string>",
    "metadata": {},
    "style": "<string>",
    "stylize": 123,
    "chaos": 123,
    "weird": 123,
    "tile": True,
    "iw": 123,
    "cref": "<string>",
    "cw": 123,
    "sref": "<string>",
    "sw": 123,
    "dref": "<string>",
    "dw": 123,
    "raw": True,
    "draft": True,
    "hd": True,
    "stop": 123,
    "extra": "<string>"
}
headers = {"Content-Type": "application/json"}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {'Content-Type': 'application/json'},
  body: JSON.stringify({
    model: '<string>',
    action: '<string>',
    prompt: '<string>',
    aspect_ratio: '<string>',
    image_urls: ['<string>'],
    seed: 123,
    negative_prompt: '<string>',
    speed: '<string>',
    version: '<string>',
    ref_task_id: '<string>',
    index: 123,
    custom_id: '<string>',
    zoom_ratio: 123,
    direction: '<string>',
    mask_url: '<string>',
    metadata: {},
    style: '<string>',
    stylize: 123,
    chaos: 123,
    weird: 123,
    tile: true,
    iw: 123,
    cref: '<string>',
    cw: 123,
    sref: '<string>',
    sw: 123,
    dref: '<string>',
    dw: 123,
    raw: true,
    draft: true,
    hd: true,
    stop: 123,
    extra: '<string>'
  })
};

fetch('https://api.qingbo.ai/v1/tasks', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.qingbo.ai/v1/tasks"

	payload := strings.NewReader("{\n  \"model\": \"<string>\",\n  \"action\": \"<string>\",\n  \"prompt\": \"<string>\",\n  \"aspect_ratio\": \"<string>\",\n  \"image_urls\": [\n    \"<string>\"\n  ],\n  \"seed\": 123,\n  \"negative_prompt\": \"<string>\",\n  \"speed\": \"<string>\",\n  \"version\": \"<string>\",\n  \"ref_task_id\": \"<string>\",\n  \"index\": 123,\n  \"custom_id\": \"<string>\",\n  \"zoom_ratio\": 123,\n  \"direction\": \"<string>\",\n  \"mask_url\": \"<string>\",\n  \"metadata\": {},\n  \"style\": \"<string>\",\n  \"stylize\": 123,\n  \"chaos\": 123,\n  \"weird\": 123,\n  \"tile\": true,\n  \"iw\": 123,\n  \"cref\": \"<string>\",\n  \"cw\": 123,\n  \"sref\": \"<string>\",\n  \"sw\": 123,\n  \"dref\": \"<string>\",\n  \"dw\": 123,\n  \"raw\": true,\n  \"draft\": true,\n  \"hd\": true,\n  \"stop\": 123,\n  \"extra\": \"<string>\"\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.qingbo.ai/v1/tasks")
  .header("Content-Type", "application/json")
  .body("{\n  \"model\": \"<string>\",\n  \"action\": \"<string>\",\n  \"prompt\": \"<string>\",\n  \"aspect_ratio\": \"<string>\",\n  \"image_urls\": [\n    \"<string>\"\n  ],\n  \"seed\": 123,\n  \"negative_prompt\": \"<string>\",\n  \"speed\": \"<string>\",\n  \"version\": \"<string>\",\n  \"ref_task_id\": \"<string>\",\n  \"index\": 123,\n  \"custom_id\": \"<string>\",\n  \"zoom_ratio\": 123,\n  \"direction\": \"<string>\",\n  \"mask_url\": \"<string>\",\n  \"metadata\": {},\n  \"style\": \"<string>\",\n  \"stylize\": 123,\n  \"chaos\": 123,\n  \"weird\": 123,\n  \"tile\": true,\n  \"iw\": 123,\n  \"cref\": \"<string>\",\n  \"cw\": 123,\n  \"sref\": \"<string>\",\n  \"sw\": 123,\n  \"dref\": \"<string>\",\n  \"dw\": 123,\n  \"raw\": true,\n  \"draft\": true,\n  \"hd\": true,\n  \"stop\": 123,\n  \"extra\": \"<string>\"\n}")
  .asString();

Text Generation

Image Generation

Video Generation

Audio

Tools & Embeddings

Task Management

Account & Usage

Quick start

Request parameters

Limits

Pricing

Response

Examples

​Quick start

​Request parameters

​Limits

​Pricing

​Response

​Examples

​Related

Quick start

Request parameters

Limits

Pricing

Response

Examples

Related