Midjourney API

Premium AI image generation with text-to-image, image-to-image, image-edit, variation, upscale, reroll, and enhance workflows

Overview

Midjourney is an async image model for high-quality diffusion generation, source-image-guided generation, region editing, variation, upscaling, reroll, and draft enhancement. Use this page when you are ready to call the API after trying Midjourney in the APIXO playground.

Capability	Value
Model ID	midjourney
Modes	text-to-image, image-to-image, image-edit, vary, upscale, reroll, enhance
Prompt length	Non-empty. Route limits are up to 2000 characters or up to 7500 UTF-8 bytes; keep prompts under 2000 characters for portable requests.
Reference images	text-to-image supports up to 20 image_urls as image prompts. image-to-image requires at least one image reference field and still uses diffusion. image-edit requires exactly 1 source image for the edit base.
Aspect ratios	1:1, 1:2, 2:1, 2:3, 3:2, 3:4, 4:3, 5:6, 6:5, 9:16, 16:9
Speeds	fast, turbo, draft; relaxed is deprecated by the official provider. To maintain compatibility, we still accept this value, but it will be converted to fast for execution and billing. draft belongs to the diffusion family, so it is available for text-to-image and image-to-image when version=7.
Result count	Usually 4 images for generation and variation, usually 1 image for upscale and enhance

Workflow categories

Category	Modes	Description
Direct generation	text-to-image, image-to-image, image-edit	Submit a new task directly. image-to-image is the diffusion route with at least one required image reference. image-edit is the region editing workflow.
Secondary editing	vary, upscale, reroll, enhance	Operate on a previous Midjourney task result. reroll uses taskId. vary, upscale, and enhance use taskId plus imageIndex. enhance is only available for draft results.

Do not put Midjourney inline flags such as --ar, --style, or --sref inside prompt. The backend strips prompt text from the first -- onward. Use API fields such as aspect_ratio, stylization, variety, and weirdness instead.

Endpoint and authentication

Base URL:

https://api.apixo.ai/api/v1

Method	Endpoint	Purpose
POST	/generateTask/midjourney	Submit a Midjourney task
GET	/statusTask/midjourney?taskId={taskId}	Poll task status and retrieve results

All requests require your APIXO API key:

Authorization: Bearer YOUR_API_KEY

Submit requests also require:

Content-Type: application/json

Copy-paste async quickstart

This minimal request submits a text-to-image task and returns a taskId.

curl -X POST "https://api.apixo.ai/api/v1/generateTask/midjourney" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "request_type": "async",
    "input": {
      "mode": "text-to-image",
      "prompt": "a cinematic product photo of a transparent speaker on black marble",
      "version": "7",
      "speed": "fast",
      "aspect_ratio": "16:9",
      "image_urls": [
        "https://example.com/scene-reference.jpg"
      ],
      "sref_urls": [
        {
          "url": "https://example.com/style-reference.jpg",
          "weight": 2
        }
      ],
      "iw": 1.5,
      "sw": 500,
      "sv": 6,
      "negative_prompt": "text, watermark"
    }
  }'

Successful response:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678"
  }
}

Save the taskId; you need it to poll for the final result.

Poll for result

curl -X GET "https://api.apixo.ai/api/v1/statusTask/midjourney?taskId=task_12345678" \
  -H "Authorization: Bearer YOUR_API_KEY"

Processing response:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "processing",
    "createTime": 1767965610929
  }
}

Success response:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "success",
    "resultJson": "{\"resultUrls\":[\"https://file.apixo.ai/img1.png\",\"https://file.apixo.ai/img2.png\",\"https://file.apixo.ai/img3.png\",\"https://file.apixo.ai/img4.png\"]}",
    "supported_actions": [
      {
        "mode": "reroll"
      },
      {
        "mode": "vary",
        "imageIndex_options": [0, 1, 2, 3],
        "type_options": [0, 1]
      },
      {
        "mode": "upscale",
        "imageIndex_options": [0, 1, 2, 3],
        "type_options": [0, 1, 2, 3]
      },
      {
        "mode": "image-edit",
        "imageIndex_options": [0, 1, 2, 3]
      },
      {
        "mode": "enhance",
        "imageIndex_options": [0, 1, 2, 3]
      }
    ],
    "createTime": 1767965610929,
    "completeTime": 1767965652317,
    "costTime": 41388
  }
}

Failed response:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "failed",
    "failCode": "PromptInvalid",
    "failMsg": "Prompt is invalid or rejected by provider",
    "createTime": 1767965610929,
    "completeTime": 1767965620132,
    "costTime": 9203
  }
}

Parse resultJson after state becomes success:

const payload = JSON.parse(data.resultJson);
const imageUrls = payload.resultUrls;

Use data.supported_actions directly when you want to present follow-up actions to users. It is returned at the same level as resultJson and is only present when follow-up operations are available.

The follow-up action set depends on the source task mode:

Source mode	Follow-up modes
text-to-image	reroll, vary, upscale, image-edit, enhance
image-to-image	reroll, vary, upscale, image-edit, enhance
image-edit	image-edit, upscale
vary	reroll, vary, upscale, image-edit
upscale	vary, image-edit
reroll	reroll, vary, upscale, image-edit, enhance
enhance	vary, upscale, image-edit

enhance is only returned for successful v7 + draft generation-family results. Follow-up actions that continue from a specific image can include imageIndex_options so clients can ask the user which image to continue from.

Request body

Text-to-image

{
  "request_type": "async",
  "input": {
    "mode": "text-to-image",
    "prompt": "a quiet library inside a glass greenhouse, soft morning light",
    "version": "7",
    "speed": "fast",
    "aspect_ratio": "16:9",
    "image_urls": [
      "https://example.com/scene-reference.jpg"
    ],
    "sref_urls": [
      {
        "url": "https://example.com/style-reference.jpg",
        "weight": 2
      }
    ],
    "iw": 1.5,
    "sw": 500,
    "sv": 6,
    "negative_prompt": "text, watermark",
    "stylization": 100,
    "variety": 10,
    "weirdness": 0
  }
}

Image-to-image

{
  "request_type": "async",
  "input": {
    "mode": "image-to-image",
    "prompt": "a dog lying in a bright park with green grass and trees",
    "image_urls": [
      "https://example.com/dog-at-home.jpg"
    ],
    "version": "7",
    "speed": "fast",
    "aspect_ratio": "4:3"
  }
}

Image-edit

Use image-edit when you want to edit a specific region. If mask, canvas, and imgPos are omitted on compatible routes, the backend can auto-fill a full-image edit region from image_urls[0].

{
  "request_type": "async",
  "input": {
    "mode": "image-edit",
    "prompt": "add a modern espresso machine on the counter",
    "image_urls": [
      "https://example.com/cafe.jpg"
    ],
    "version": "7",
    "speed": "fast",
    "mask": {
      "areas": [
        {
          "width": 1024,
          "height": 1024,
          "points": [320, 420, 700, 420, 700, 760, 320, 760]
        }
      ]
    },
    "canvas": {
      "width": 1024,
      "height": 1024
    },
    "imgPos": {
      "width": 1024,
      "height": 1024,
      "x": 0,
      "y": 0
    }
  }
}

Vary

{
  "request_type": "async",
  "input": {
    "mode": "vary",
    "taskId": "task_previous_midjourney",
    "imageIndex": 0,
    "type": 1,
    "remixPrompt": "more dramatic lighting and richer color contrast"
  }
}

Upscale

{
  "request_type": "async",
  "input": {
    "mode": "upscale",
    "taskId": "task_previous_midjourney",
    "imageIndex": 2,
    "type": 0
  }
}

Reroll

{
  "request_type": "async",
  "input": {
    "mode": "reroll",
    "taskId": "task_previous_midjourney"
  }
}

Enhance

{
  "request_type": "async",
  "input": {
    "mode": "enhance",
    "taskId": "task_previous_midjourney",
    "imageIndex": 1
  }
}

Parameters

request_type · string · default: "async" · required

Result delivery mode. Use async for polling with statusTask, or callback for webhook delivery.

callback_url · string

Required when request_type is callback. Must be a public HTTPS URL that can receive the final task payload. See Webhooks.

input · object · required

Midjourney input parameters.

Show properties

mode · string · required

Operation mode. Supported values: text-to-image, image-to-image, image-edit, vary, upscale, reroll, enhance.

prompt · string

Required for text-to-image, image-to-image, and image-edit. Must be non-empty. Route limits are up to 2000 characters or up to 7500 UTF-8 bytes; keep prompts under 2000 characters for portable requests.

image_urls · string[]

Reference image URLs. For text-to-image, this field is optional and supports up to 20 image prompts. For image-to-image, at least one image reference field is required across image_urls, sref_urls, cref_urls, or oref_urls; if image_urls is used, it works as the direct image prompt input. For image-edit, it is required and must contain exactly 1 base image. URLs should be public, directly fetchable JPG, PNG, or WebP files.

sref_urls · object[]

Optional style reference images for text-to-image. Each item must contain url and can optionally contain weight. weight is a relative per-image weight, not a fixed-sum ratio.

cref_urls · string[]

Optional character reference images for text-to-image.

oref_urls · string[]

Optional object reference images for text-to-image. At most 1 image is allowed.

version · string

Midjourney model version. Use 7, 6.1, 6, 8.1, niji6, or niji7 for current compatibility. The API also normalizes forms such as v7, v6.1, v6, v8.1, niji 6, and niji 7.

speed · string

Required for text-to-image, image-to-image, and image-edit. Supported values: relaxed, fast, turbo, and draft. relaxed is deprecated by the official provider. To maintain compatibility, APIXO still accepts it, but it is converted to fast for execution and billing. draft is part of the diffusion family, so it is available for text-to-image and image-to-image when version=7. image-edit does not support draft.

aspect_ratio · string

Output aspect ratio. Supported values: 1:1, 1:2, 2:1, 2:3, 3:2, 3:4, 4:3, 5:6, 6:5, 9:16, 16:9. When omitted, the upstream default is used.

variety · integer

Midjourney chaos/variation control. Supported range: 0-100. One route defaults omitted values to 10; other routes omit the flag unless you set it.

stylization · integer

Artistic stylization strength. Supported range: 0-1000. One route defaults omitted values to 1; other routes omit the flag unless you set it.

weirdness · integer

Surreal or unconventional effect strength. Supported range: 0-3000. One route defaults omitted values to 1; other routes omit the flag unless you set it.

iw · number

Image prompt weight for text-to-image. This field is only active when image_urls is provided.

sw · integer

Style reference weight for text-to-image. Supported range: 0-1000. This field is only active when sref_urls is provided.

cw · integer

Character reference weight for text-to-image. Supported range: 0-100. This field is only active when cref_urls is provided.

ow · integer

Object reference weight for text-to-image. Supported range: 1-1000. This field is only active when oref_urls is provided.

sv · integer

Style reference version for text-to-image. This field is only active when sref_urls is provided.

exp · integer

Experimental parameter for text-to-image. Supported range: 0-100 on supported versions.

negative_prompt · string

Negative prompt for text-to-image. This maps to Midjourney --no.

quality · number

Optional quality setting. Supported values are version-specific.

is_hd · boolean

Optional 2K HD switch. Only supported in text-to-image when version is 8.1.

is_raw · boolean

Optional raw mode switch for text-to-image. When true, the request adds Midjourney --raw.

enableTranslation · boolean

Requests prompt translation to English. Translation behavior is route-dependent: some routes default to translation, while others only translate when the route is configured for it and this flag is true.

watermark · string

Optional watermark text. Supported on compatible routes.

taskId · string

Required for vary, upscale, reroll, and enhance. Use the APIXO task ID from a previous successful Midjourney generation task.

imageIndex · integer

Required for vary, upscale, and enhance. Selects one image from the previous result. Supported range: 0-3.

type · integer · default: "0"

Optional transform type. For vary, 0 means subtle variation and 1 means strong variation. For upscale, supported values are 0, 1, 2, and 3 on compatible routes.

remixPrompt · string

Optional text guidance for vary on compatible routes. For image-edit, if remixPrompt is omitted, compatible routes reuse prompt.

mask · object

Optional for image-edit. Use it when you need region-specific editing. If omitted on compatible routes, the backend can create a full-image mask from the source image.

Show properties

areas · object[]

Polygon edit regions. Each area can include width, height, and points. points is a flat coordinate array: [x1,y1,x2,y2,...].

url · string

Black/white mask image URL or base64 mask. White areas are regenerated. Use either areas or url, not both.

canvas · object

Optional for image-edit. Canvas dimensions. If provided, width and height must be positive integers.

Show properties

width · integer

Canvas width in pixels.

height · integer

Canvas height in pixels.

imgPos · object

Optional for image-edit. Source image position and size on the canvas.

Show properties

width · integer

Image width in pixels.

height · integer

Image height in pixels.

x · integer

Horizontal offset from the canvas top-left corner.

y · integer

Vertical offset from the canvas top-left corner.

text-to-image and image-to-image share the diffusion parameter family. image-to-image requires at least one image reference across image_urls, sref_urls, cref_urls, or oref_urls, while text-to-image treats image_urls as optional image prompts. image-edit is a separate region-editing workflow and uses mask, canvas, imgPos, and optional remixPrompt.

Response format

Submit task response

POST /generateTask/midjourney returns a task ID when the task is accepted:

code · integer

API status code. 200 means the task was accepted.

message · string

Human-readable status message.

data.taskId · string

Unique APIXO task identifier used with the status endpoint.

Status response fields

taskId · string

Unique APIXO task identifier.

state · string

Current task state: pending, processing, success, or failed.

resultJson · string

JSON string containing the generated image URLs. Present when state is success.

supported_actions · array

Optional follow-up actions for successful results. This field is returned at the same level as resultJson in data.

failCode · string

Machine-readable failure code. Present when state is failed.

failMsg · string

Human-readable failure message. Present when state is failed.

createTime · integer

Task creation timestamp in Unix milliseconds.

completeTime · integer

Task completion timestamp in Unix milliseconds. Present after completion.

costTime · integer

Processing duration in milliseconds. Present after completion.

Webhook callback mode

Use callback mode when your backend should receive the final result automatically instead of polling.

curl -X POST "https://api.apixo.ai/api/v1/generateTask/midjourney" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "request_type": "callback",
    "callback_url": "https://your-server.com/webhooks/apixo",
    "input": {
      "mode": "text-to-image",
      "prompt": "a premium watch on dark stone, macro photography, soft reflections",
      "version": "7",
      "speed": "fast",
      "aspect_ratio": "1:1"
    }
  }'

See Webhooks for delivery requirements and retry behavior.

Billing

Midjourney is billed per use. The selected mode and speed determine the unit price.

The current public pricing is easiest to understand in two families:

Category	Mode	APIXO price	Notes
Direct generation	text-to-image	$0.10 / use (draft), $0.20 / use (fast), $0.40 / use (turbo)	Diffusion generation. v8.1 with is_hd=true and fast is $0.30 / use.
Direct generation	image-to-image	Same as text-to-image	Uses the same diffusion route as text-to-image, but requires at least one image reference input.
Direct generation	image-edit	$0.30 / use (fast), $0.60 / use (turbo)	Region editing workflow.
Secondary editing	vary	$0.30 / use	Requires a previous taskId and imageIndex.
Secondary editing	upscale	$0.30 / use	Requires a previous taskId and imageIndex.
Secondary editing	reroll	Same as the source task price	Re-runs the previous successful task with the same configuration.
Secondary editing	enhance	See Pricing	Draft-only enhancement workflow. Requires a previous taskId and imageIndex.

For current route and market comparison pricing, see Pricing.

Latency and polling

Actual latency may vary by prompt complexity, mode, selected route, and current queue load.

Mode or speed	Typical generation time	Recommended first poll	Poll interval
turbo generation	15s-25s	20s after task creation	3s-5s
fast generation	30s-45s	20s after task creation	3s-5s
vary, upscale, or enhance	15s-60s	15s after task creation	3s-5s
reroll	Similar to the original generation speed	20s after task creation	3s-5s

For high-concurrency production workloads, use callback mode to avoid frequent polling.

Rate limits and concurrency can vary by account, API key, and route. If you receive 429, slow down requests and retry with backoff. For account-level details, see System APIs.

Errors and troubleshooting

HTTP errors

Code	Meaning	What to do
400	Invalid request body, mode, parameter, prompt, or image URL	Fix the request before retrying
401	Missing or invalid API key	Check the Authorization header
402	Insufficient balance or quota	Add balance or switch account/key
403	Key or route cannot access the model	Check permissions and route strategy
429	Rate limit or concurrency limit reached	Retry with exponential backoff
500	Server error	Retry with backoff
502	Upstream provider error	Retry with backoff
504	Upstream timeout	Retry or use callback mode for long-running jobs

Task failure codes

Fail code	Meaning	What to do
PromptInvalid	Prompt is invalid or rejected by the provider	Rewrite the prompt and remove unsafe or malformed content
MissingParameter	Required request parameters are missing	Check mode, prompt, speed, image_urls, taskId, and imageIndex
SensitiveContentDetected	Input or output was flagged by safety checks	Change the prompt or source image
InputOutputSensitiveContentDetected	The input or output was flagged as sensitive	Try different inputs
Upload error	Image upload or image size validation failed	Use a public direct image URL and keep image files within supported limits
Website TimeOut	MidJourney official website did not respond after multiple attempts	Retry later or use callback mode
Task TimeOut	The upstream task became outdated or timed out	Retry the generation
RateLimitExceeded	Upstream or APIXO rate limit was reached	Retry with exponential backoff

Common fixes

Symptom	Fix
Invalid mode type	Use one of text-to-image, image-to-image, image-edit, vary, upscale, reroll, or enhance.
Parameter is ignored	Use stable API field names: aspect_ratio, variety, stylization, and weirdness.
Prompt unexpectedly loses flags	Move Midjourney flags out of prompt; text after the first -- is stripped by the backend.
taskId not found for vary, upscale, reroll, or enhance	Use the APIXO task ID from a previous Midjourney result, not the upstream provider task ID.
Image-edit fails to read image size	Use a public, directly fetchable source image URL. The backend reads image_urls[0] to fill default canvas and mask values.
relaxed does not match expected execution speed or billing	relaxed is accepted for compatibility, but APIXO converts it to fast for execution and billing.
draft is rejected	draft is only supported in the diffusion family (text-to-image and image-to-image) when version=7.
enhance is rejected	enhance only works for successful Midjourney draft results.

See Error Codes for the full error reference.

Related links

• Generation API Overview
• Image Models
• Generate Task
• Status Task
• Webhooks
• Error Codes
• Parameter Specification
• Routing Strategies
• Try Midjourney in the APIXO Playground
• Pricing