Advanced

Advanced Techniques

Master Midjourney's advanced features: blend, describe, remix mode, pan, zoom, region variation, and permutation prompts.

/blend

Blend combines 2-5 images into a single new image, merging their concepts, styles, and compositions:

Blend Command
/blend
  image1: [upload sunset photo]
  image2: [upload mountain landscape]
  # Creates a blend of both images

# Optional: set dimensions
  dimensions: Portrait, Square, or Landscape

# You can blend up to 5 images
# Great for style mixing and concept fusion

/describe

Upload an image and Midjourney will generate 4 text prompts that could recreate a similar image. This is invaluable for reverse-engineering prompts and learning what language Midjourney responds to:

Describe Command
/describe
  image: [upload any image]

# Midjourney returns 4 prompt suggestions like:
1. a woman standing in a field of sunflowers, golden hour,
   warm tones, editorial photography, depth of field
2. portrait in sunflower field, natural light, soft focus,
   fashion photography, summer vibes
3. ...
4. ...

# Click the numbered buttons to run any suggestion as a prompt

Remix Mode

When remix is enabled, clicking Variation buttons lets you edit the prompt before generating. This allows you to change the subject, style, or parameters while keeping the composition:

Remix
# Enable remix mode
/settings -> Toggle "Remix mode" ON

# Or use the command
/prefer remix

# Now when you click V1-V4, a prompt editor appears
# Original: "a cat sitting on a windowsill"
# Remix to: "a dog sitting on a windowsill, oil painting"
# Same composition, different subject and style!

Pan & Zoom

After upscaling an image, you can extend it:

  • Pan: Click the arrow buttons to extend the image in any direction (left, right, up, down). The AI generates new content that seamlessly continues the existing image.
  • Zoom Out (1.5x/2x): Expands the canvas and fills the new space while keeping the original image centered. Great for revealing more of a scene.
  • Custom Zoom: Click "Custom Zoom" to set a specific zoom level and optionally modify the prompt for the expanded area.

Region Variation

Select specific regions of an upscaled image to regenerate while keeping the rest intact. This is Midjourney's version of inpainting:

  1. Upscale an image (U1-U4)
  2. Click "Vary (Region)"
  3. Draw a selection around the area you want to change
  4. Optionally modify the prompt for the selected region
  5. Submit to regenerate just that area

Permutation Prompts

Generate multiple variations from a single prompt using curly braces:

Permutations
# This creates 3 separate jobs:
/imagine a {red, blue, green} sports car on a mountain road
# Job 1: a red sports car on a mountain road
# Job 2: a blue sports car on a mountain road
# Job 3: a green sports car on a mountain road

# Combine with parameters:
/imagine landscape --ar {1:1, 16:9, 9:16}

# Multiple permutations (creates all combinations):
/imagine a {cat, dog} in a {forest, city} --s {100, 500}
# Creates 2 x 2 x 2 = 8 separate jobs
Caution: Permutation prompts can consume your generation quota quickly. A prompt with 3 options in each of 3 brackets creates 27 separate jobs. Use mindfully.

What's Next?

The final lesson covers best practices for professional workflows, commercial use, ethical considerations, and prompt optimization strategies.