Midjourney

Premium AI image generation with text-to-image, image-to-image, upscale, and variation

Endpoints

Authentication

All requests require an API Key in the header:

Authorization: Bearer YOUR_API_KEY

Request Body

{
  "request_type": "async",
  "callback_url": "https://...",
  "provider": "auto",
  "input": {
    "mode": "text-to-image",
    "version": "v7",
    "speed": "fast",
    "prompt": "..."
  }
}

Parameters

Common Parameters

Input Parameters

All modes require the input object with mode-specific parameters.

Universal Input Parameters:

Optional Enhancement Parameters:

Model Version Guide

Available Versions: v6, v6.1, v7, niji 6, niji 7 (Default: v6.1)

v7 (Latest, April 2025)

• Best overall quality with breakthrough improvements in text/image understanding
• Exceptional accuracy in human anatomy, hand details, and object textures
• Includes Sketch Mode for rapid iteration and enhanced image reference capabilities
• Recommended for: Professional work requiring highest quality and detail

v6.1 (Default, July 2024)

• ~25% faster than v6 with more coherent results
• Precise details and textures with excellent balance of speed and quality
• Recommended for: General use and production workflows

v6 (December 2023)

• Improved prompt accuracy for longer inputs and enhanced coherence
• Advanced image prompting and remix capabilities
• Recommended for: Projects requiring specific v6 characteristics

niji 7 (Latest for Anime)

• Professional anime creation with precise character control and rich stylization
• Dramatic lighting, flexible style adjustment from photorealistic to highly stylized
• Superior multi-character consistency and aesthetic quality
• Recommended for: Anime, manga, and illustration projects

niji 6 (June 2024)

• Specialized for Eastern/anime aesthetics with improved Japanese text rendering
• Better anime eye structure and detail rendering
• Recommended for: Anime-style illustrations with Japanese text

Mode: text-to-image (Text-to-Image Generation)

Generate images from text descriptions.

Note: text-to-image mode also supports optional enhancement parameters (aspect, chaos, stylize, weird, enableTranslation) listed above.

Mode: image-to-image (Image-to-Image Generation)

Generate new images based on reference images. Use descriptive language in prompts rather than editing instructions.

Best Practice: Describe the desired result, not the editing action.

• ✅ Good: "a dog lying in a park"
• ❌ Bad: "change the background to a park"

Mode: upload-paint (Inpainting / Region Editing)

Edit specific regions of an image by painting a mask. Requires precise mask definition.

Mode: vary (Create Variations)

Generate similar but varied versions of a previous result.

Mode: upscale (High-Resolution Upscaling)

Upscale a previous result to higher resolution.

Available Modes

Currently Available:

• text-to-image — Text-to-image generation
• image-to-image — Image-to-image generation with reference images
• upload-paint — Inpaint specific regions (requires mask)
• vary — Create variations with similar style
• upscale — High-resolution upscaling

Additional Capabilities (Contact us to enable):

• Outpaint — Extend image boundaries
• Retexture — Apply new textures to images
• Remove Background — Background removal
• Reroll — Re-execute generation task

Examples

Text-to-Image

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/callback",
    "provider": "value",
    "input": {
      "mode": "text-to-image",
      "prompt": "一个温馨的赛博朋克咖啡馆，霓虹灯光，电影级照明",
      "version": "v7",
      "speed": "fast",
      "aspect": "16:9",
      "enableTranslation": true,
      "stylize": 500,
      "chaos": 20
    }
  }'

Image-to-Image

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",
    "provider": "official",
    "input": {
      "mode": "image-to-image",
      "prompt": "a dog lying in a beautiful park with green grass and trees",
      "image_urls": ["https://example.com/dog-at-home.jpg"],
      "version": "v7",
      "speed": "fast"
    }
  }'

Upload-Paint (Inpainting)

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",
    "provider": "official",
    "input": {
      "mode": "upload-paint",
      "prompt": "add a modern coffee machine in the center",
      "image_urls": ["https://example.com/cafe.jpg"],
      "version": "v7",
      "speed": "fast",
      "mask": {
        "areas": [{
          "width": 512,
          "height": 512,
          "points": [10,10,10,100,100,100,100,10]
        }]
      },
      "canvas": {
        "width": 1024,
        "height": 1024
      },
      "imgPos": {
        "width": 512,
        "height": 512,
        "x": 256,
        "y": 256
      }
    }
  }'

Vary (Create Variation)

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": "vary",
      "taskId": "task_previous_midjourney",
      "imageIndex": 0,
      "type": 1,
      "prompt": "add more vibrant colors"
    }
  }'

Upscale

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": "upscale",
      "taskId": "task_previous_midjourney",
      "imageIndex": 2,
      "type": 1
    }
  }'

Response

POST /api/v1/generateTask/midjourney

Returns taskId on success for subsequent status queries.

Success:

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

Error:

{
  "code": 400,
  "message": "Insufficient credits",
  "data": null
}

GET /api/v1/statusTask/midjourney

Query task execution status and results via taskId.

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

Success:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "success",
    "resultJson": "{\"resultUrls\":[\"https://r2.apixo.ai/img1.png\",\"https://r2.apixo.ai/img2.png\",\"https://r2.apixo.ai/img3.png\",\"https://r2.apixo.ai/img4.png\"]}",
    "createTime": 1767965610929,
    "completeTime": 1767965652317,
    "costTime": 41388
  }
}

Failed:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "failed",
    "failCode": "CONTENT_VIOLATION",
    "failMsg": "Content does not meet safety guidelines"
  }
}

Status Response Fields

Error Codes

Rate Limits

Exceeding limits returns 429 error. Wait and retry.

Tips

• Prompt translation: Enable enableTranslation: true for non-English prompts. English typically yields better results in Midjourney.

• Aspect ratio: Choose based on use case:

  • 1:1 - Social media posts, avatars
  • 4:3 - Classic photography, presentations
  • 16:9 - Widescreen, videos, banners
  • 9:16 - Mobile stories, vertical videos

• Chaos parameter: Control result diversity:

  • 0-30 - Consistent, predictable results (recommended for production)
  • 30-70 - Balanced variety with creative exploration
  • 70-100 - Maximum diversity, experimental outputs

• Stylize parameter: Balance realism vs artistry:

  • 0-250 - Literal interpretation, realistic
  • 250-750 - Balanced (default: 100)
  • 750-1000 - Highly artistic, creative interpretation

• Weird parameter: Add surreal/unconventional elements:

  • 10-300 - Subtle unusual touches
  • 300-1000 - Noticeable non-traditional elements
  • 1000-3000 - Extreme surrealism and abstraction

• Generation time:

  • fast mode: ~30-45 seconds, generates 4 images
  • turbo mode: ~15-25 seconds, faster
  • Submit task, wait 20 seconds, then poll every 3 seconds

• Callback mode: Use callback mode to avoid frequent polling.

• Image expiration: Result URLs are valid for 15 days. Download promptly.

• Content moderation: Prompts must comply with safety guidelines.

• Version selection: See the Model Version Guide above for detailed recommendations on choosing the right version for your use case

• Recommended workflow:

  1. Use text-to-image to generate initial 4 images
  2. Select preferred image (0-3), use upscale for high-resolution output
  3. Or use vary to create more variations
  4. Iterate until desired result

• Image-to-image mode:

  • Use descriptive language in prompts: describe what you want to see, not editing actions
  • Example: "a dog in a park" instead of "change background to park"
  • More reference images = less control over output. Use 1-2 images for best results

• Upload-paint mode (inpainting):

  • mask.areas: Define polygon regions with width, height, and points array. Points are in clockwise order as [x1,y1,x2,y2,...] from top-left origin
  • mask.url: Black/white mask image with same dimensions as original (white = regenerate region). Supports base64 encoding
  • imgPos: Specifies image position and size on canvas

• Upscale types:

  • Type 0: Standard HD
  • Type 1: Creative HD (adds creative details)
  • Type 2: 2x HD resolution
  • Type 3: 4x HD resolution

• Image formats: image_urls supports JPG, PNG, WebP, max 10MB per image.

Related

• Pricing
• Error Codes