Midjourney - QWave API

POST

tasks

Midjourney

curl --request POST \
  --url https://www.qingbo.dev/v1/tasks \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "action": "<string>",
  "prompt": "<string>",
  "aspect_ratio": "<string>",
  "image_urls": [
    "<string>"
  ],
  "seed": 123,
  "callback_url": "<string>",
  "speed": "<string>",
  "ref_task_id": "<string>",
  "index": 123,
  "direction": "<string>",
  "zoom_ratio": 123,
  "custom_id": "<string>",
  "version": "<string>",
  "niji": true,
  "quality": "<string>",
  "stylize": 123,
  "chaos": 123,
  "negative_prompt": "<string>"
}
'

import requests

url = "https://www.qingbo.dev/v1/tasks"

payload = {
    "model": "<string>",
    "action": "<string>",
    "prompt": "<string>",
    "aspect_ratio": "<string>",
    "image_urls": ["<string>"],
    "seed": 123,
    "callback_url": "<string>",
    "speed": "<string>",
    "ref_task_id": "<string>",
    "index": 123,
    "direction": "<string>",
    "zoom_ratio": 123,
    "custom_id": "<string>",
    "version": "<string>",
    "niji": True,
    "quality": "<string>",
    "stylize": 123,
    "chaos": 123,
    "negative_prompt": "<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,
    callback_url: '<string>',
    speed: '<string>',
    ref_task_id: '<string>',
    index: 123,
    direction: '<string>',
    zoom_ratio: 123,
    custom_id: '<string>',
    version: '<string>',
    niji: true,
    quality: '<string>',
    stylize: 123,
    chaos: 123,
    negative_prompt: '<string>'
  })
};

fetch('https://www.qingbo.dev/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://www.qingbo.dev/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  \"callback_url\": \"<string>\",\n  \"speed\": \"<string>\",\n  \"ref_task_id\": \"<string>\",\n  \"index\": 123,\n  \"direction\": \"<string>\",\n  \"zoom_ratio\": 123,\n  \"custom_id\": \"<string>\",\n  \"version\": \"<string>\",\n  \"niji\": true,\n  \"quality\": \"<string>\",\n  \"stylize\": 123,\n  \"chaos\": 123,\n  \"negative_prompt\": \"<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://www.qingbo.dev/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  \"callback_url\": \"<string>\",\n  \"speed\": \"<string>\",\n  \"ref_task_id\": \"<string>\",\n  \"index\": 123,\n  \"direction\": \"<string>\",\n  \"zoom_ratio\": 123,\n  \"custom_id\": \"<string>\",\n  \"version\": \"<string>\",\n  \"niji\": true,\n  \"quality\": \"<string>\",\n  \"stylize\": 123,\n  \"chaos\": 123,\n  \"negative_prompt\": \"<string>\"\n}")
  .asString();

{
  "task_id": "task-wave1781770984b950327128",
  "model": "midjourney",
  "action": "imagine",
  "status": "queued",
  "created_at": 1781770984120,
  "progress": 0
}

Midjourney 通过统一异步任务接口 /v1/tasks 调用,用 action 字段区分操作,一个 model 暴露全部能力。 两类操作:

独立生成类 — 不依赖前序任务:imagine(文生图)、blend(多图融合)、describe(图转文)、edits(全图重写)
链式操作类 — 基于前一个任务的产物再加工,需传 ref_task_id 引用前序任务:upscale(放大)、variation / high-variation / low-variation(变体)、reroll(重出网格)、zoom(外绘缩放)、pan(方向平移)、remix-strong / remix-subtle(v8 重塑)

imagine 等生成类返回 2×2 四宫格(一张图含 4 个画面)。要对其中某一张放大 / 变体,用链式操作传 ref_task_id(该网格任务的 task_id)+ index(1-4 选第几张)。详见链式操作。

定价

按操作 + 速度档计费(单位 USD)。speed 取 relax(默认) / fast / turbo。

操作	relax / 默认	fast	turbo
`imagine`	`$0.05067`	`$0.06192`	`$0.1125`
其余全部 action	`$0.06192`	`$0.06192`	`$0.1125`

任务失败自动全额退款。

Action 速查

action	类型	必填	说明
`imagine`	生成	`prompt`	文生图;可加 `image_urls` 做图像引导。出四宫格
`blend`	生成	`image_urls`(2-4)	多图融合,无 prompt
`describe`	生成	`image_urls`(1)	图转文,返回 4 条提示词建议
`edits`	生成	`prompt` + `image_urls`(1)	全图重写(背景替换 / 风格迁移)
`upscale`	链式	`ref_task_id` + `index`	放大网格中第 `index` 张
`variation`	链式	`ref_task_id` + `index`	基于某张出新的四宫格变体
`high-variation`	链式	`ref_task_id` + `index`	强变体(变化幅度大)
`low-variation`	链式	`ref_task_id` + `index`	弱变体(保持主体)
`reroll`	链式	`ref_task_id`	用同 prompt 重出一组四宫格
`zoom`	链式	`ref_task_id` + `zoom_ratio`	外绘缩放(`<2` 用 1.5x,`≥2` 用 2x)
`pan`	链式	`ref_task_id` + `direction`	向 `left/right/up/down` 平移外绘
`remix-strong`	链式	`ref_task_id` + `index`	v8/v8.1 大幅重塑,可改 prompt
`remix-subtle`	链式	`ref_task_id` + `index`	v8/v8.1 保守重塑

调用示例

curl -X POST https://www.qingbo.dev/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "imagine",
    "prompt": "a majestic eagle soaring over snowy mountains, cinematic",
    "aspect_ratio": "16:9",
    "version": "7",
    "speed": "fast"
  }'

curl -X POST https://www.qingbo.dev/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "upscale",
    "ref_task_id": "task-wave1781770984b950327128",
    "index": 2
  }'

curl -X POST https://www.qingbo.dev/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "blend",
    "image_urls": [
      "https://cdn.example.com/a.jpg",
      "https://cdn.example.com/b.jpg"
    ],
    "aspect_ratio": "1:1"
  }'

curl -X POST https://www.qingbo.dev/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "describe",
    "image_urls": ["https://cdn.example.com/photo.jpg"]
  }'

curl -X POST https://www.qingbo.dev/v1/tasks \
  -H "Authorization: Bearer $WAVE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "midjourney",
    "action": "pan",
    "ref_task_id": "task-wave...upscale_task",
    "direction": "right"
  }'

{
  "task_id": "task-wave1781770984b950327128",
  "model": "midjourney",
  "action": "imagine",
  "status": "queued",
  "created_at": 1781770984120,
  "progress": 0
}

提交后用 GET /v1/tasks/{task_id} 轮询状态,详见任务系统。

可用模型

模型 ID	说明
`midjourney`	Midjourney 全功能,15 种 action(生成 + 链式),支持 v5.1–v8.1 / niji

链式操作

链式 action 基于前一个任务的产物再加工,核心是 ref_task_id 和 index。

先生成网格

调 imagine 拿到响应里的 task_id(例如 task-wave1781770984b950327128),它对应一张 2×2 四宫格。

引用网格 + 选图

链式操作传 ref_task_id(= 上一步的 task_id)+ index(1-4,左上=1 右上=2 左下=3 右下=4)。例如 upscale 放大第 2 张。

继续链式

放大后的单图任务又可作为 zoom / pan 的 ref_task_id,层层叠加。

ref_task_id 必须是你自己、已成功、同一渠道的任务,否则报错(source_task_not_found / source_task_not_owned / source_task_not_success / source_task_channel_mismatch)。引用的是任务 task_id,不是结果图片 URL。

各链式 action 对来源任务的要求:

action	来源任务	选图
`upscale` / `variation`	imagine / reroll 的四宫格	`index` 必填
`high-variation` / `low-variation`	upscale 后的单图	`index` 必填
`reroll`	imagine 的四宫格	不需 index
`zoom` / `pan`	upscale 后的单图	不需 index
`remix-strong` / `remix-subtle`	v8/v8.1 的 imagine 四宫格	`index` 必填

通用参数

model

string

必填

固定 midjourney

action

string

默认值:"generate"

操作类型,见 Action 速查。生成类传 imagine/blend/describe/edits,链式类传对应 action

prompt

string

提示词。imagine/edits 必填;支持原生 MJ flags(如 --ar 16:9)。blend/describe 不用

aspect_ratio

string

默认值:"1:1"

画面比例,如 1:1 / 16:9 / 9:16 / 2:3 / 3:2 / 4:3 / 3:4 / 21:9 / 9:21

image_urls

string[]

参考图。blend 2-4 张,describe/edits 1 张,imagine 可选(图像引导)

seed

integer

随机种子

callback_url

string

任务终态回调地址(可选,替代轮询)

模型特定参数

speed

string

默认值:"relax"

速度档 relax / fast / turbo,影响出图速度与计费(见定价)

ref_task_id

string

链式操作必填 — 前序任务的 task_id(本平台返回的 wave task_id)。见链式操作

index

integer

选四宫格第几张(1-4)。upscale/variation/high-variation/low-variation/remix-* 需填

direction

string

pan 平移方向:left / right / up / down

zoom_ratio

number

zoom 缩放倍数:<2 走 Outpaint 1.5x,≥2 走 CustomZoom 2x

custom_id

string

高级:直接指定 MJ 按钮 ID(给了则忽略 index),用于网关未抽象的特殊操作

version

string

MJ 版本:8.1 / 7 / 6.1 / 5.2 / 5.1

niji

boolean

切换 Niji(动漫)模式

quality

string

画质 --q:0.25 / 0.5 / 1 / 2

stylize

integer

风格化强度 --s,0-1000

chaos

integer

变异度 --c,0-100

negative_prompt

string

排除元素 --no

资源限制

blend 参考图 2-4 张,单张 ≤ 12 MiB
describe 同步返回较快(1-3 秒),结果为文本建议
pan 仅 v6/v6.1/v7/niji6 支持;v8/v8.1 已移除
remix-strong / remix-subtle 仅 v8/v8.1

相关文档

任务系统 — 异步提交 / 轮询 / 回调
图像生成概述

视频生成概述通过 /v1/tasks 统一端点调用所有视频生成模型,异步任务异步取结果

Midjourney

curl --request POST \
  --url https://www.qingbo.dev/v1/tasks \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "<string>",
  "action": "<string>",
  "prompt": "<string>",
  "aspect_ratio": "<string>",
  "image_urls": [
    "<string>"
  ],
  "seed": 123,
  "callback_url": "<string>",
  "speed": "<string>",
  "ref_task_id": "<string>",
  "index": 123,
  "direction": "<string>",
  "zoom_ratio": 123,
  "custom_id": "<string>",
  "version": "<string>",
  "niji": true,
  "quality": "<string>",
  "stylize": 123,
  "chaos": 123,
  "negative_prompt": "<string>"
}
'

import requests

url = "https://www.qingbo.dev/v1/tasks"

payload = {
    "model": "<string>",
    "action": "<string>",
    "prompt": "<string>",
    "aspect_ratio": "<string>",
    "image_urls": ["<string>"],
    "seed": 123,
    "callback_url": "<string>",
    "speed": "<string>",
    "ref_task_id": "<string>",
    "index": 123,
    "direction": "<string>",
    "zoom_ratio": 123,
    "custom_id": "<string>",
    "version": "<string>",
    "niji": True,
    "quality": "<string>",
    "stylize": 123,
    "chaos": 123,
    "negative_prompt": "<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,
    callback_url: '<string>',
    speed: '<string>',
    ref_task_id: '<string>',
    index: 123,
    direction: '<string>',
    zoom_ratio: 123,
    custom_id: '<string>',
    version: '<string>',
    niji: true,
    quality: '<string>',
    stylize: 123,
    chaos: 123,
    negative_prompt: '<string>'
  })
};

fetch('https://www.qingbo.dev/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://www.qingbo.dev/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  \"callback_url\": \"<string>\",\n  \"speed\": \"<string>\",\n  \"ref_task_id\": \"<string>\",\n  \"index\": 123,\n  \"direction\": \"<string>\",\n  \"zoom_ratio\": 123,\n  \"custom_id\": \"<string>\",\n  \"version\": \"<string>\",\n  \"niji\": true,\n  \"quality\": \"<string>\",\n  \"stylize\": 123,\n  \"chaos\": 123,\n  \"negative_prompt\": \"<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://www.qingbo.dev/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  \"callback_url\": \"<string>\",\n  \"speed\": \"<string>\",\n  \"ref_task_id\": \"<string>\",\n  \"index\": 123,\n  \"direction\": \"<string>\",\n  \"zoom_ratio\": 123,\n  \"custom_id\": \"<string>\",\n  \"version\": \"<string>\",\n  \"niji\": true,\n  \"quality\": \"<string>\",\n  \"stylize\": 123,\n  \"chaos\": 123,\n  \"negative_prompt\": \"<string>\"\n}")
  .asString();

{
  "task_id": "task-wave1781770984b950327128",
  "model": "midjourney",
  "action": "imagine",
  "status": "queued",
  "created_at": 1781770984120,
  "progress": 0
}

​定价

​Action 速查

​调用示例

​可用模型

​链式操作

​通用参数

​模型特定参数

​资源限制

​相关文档

定价

Action 速查

调用示例

可用模型

链式操作

通用参数

模型特定参数

资源限制

相关文档