Generating Consistent Illustration Sets with Midjourney

The key to consistent illustration sets in Midjourney is the --sref (Style Reference) parameter, which locks the visual style of a first generation and applies it to all subsequent ones.

First, you need a reference image. You have two options:

• Option A: Generate a reference illustration from scratch
• Option B: Use an existing illustration (from the client's brand, from a style you've agreed on, or from a reference you own)

For Option A, generate your first "master" illustration:

/imagine flat vector illustration, young woman working at laptop in cozy home office, warm pastel colors, minimal style, clean lines, white background, editorial --ar 1:1

1. After generating your first illustration, upscale your preferred version (click U1/U2/U3/U4).
2. Click the upscaled image to open it full-size.
3. Right-click → Copy Image Address (or Copy Link Address) to get the image URL.
4. You'll use this URL in the --sref parameter for all subsequent illustrations.

Use your style reference URL in every new prompt:

/imagine flat vector illustration, [new subject/scene], warm pastel colors, minimal style, clean lines, white background --ar 1:1 --sref [your-reference-image-URL] --sw 100

The --sw 100 controls style weight (0–1000) — 100 is a moderate style lock. Increase to --sw 500 for stricter style adherence, decrease to --sw 50 for looser.

Example set for a wellness brand — 6 illustrations:

1. Person meditating: flat vector illustration, woman meditating on yoga mat, morning sunlight, --ar 1:1 --sref [URL] --sw 100
2. Healthy cooking: flat vector illustration, person cooking healthy meal in bright kitchen, --ar 1:1 --sref [URL] --sw 100
3. Outdoor walk: flat vector illustration, couple walking in park, autumn trees, --ar 1:1 --sref [URL] --sw 100
4. Reading/relaxing: flat vector illustration, person reading book on comfortable sofa, warm evening light, --ar 1:1 --sref [URL] --sw 100
5. Video call: flat vector illustration, woman on video call, home office, plants in background, --ar 1:1 --sref [URL] --sw 100
6. Journaling: flat vector illustration, person writing in journal at café table, --ar 1:1 --sref [URL] --sw 100

After generating all illustrations:

1. Download all upscaled versions.
2. Place them side by side in Figma or Illustrator.
3. Check for visual consistency in: color palette range, line weight, level of detail, character proportions.
4. For any outliers, regenerate with a higher --sw value to force tighter style adherence.

For truly production-ready vector illustrations:

1. Open the Midjourney image in Adobe Illustrator.
2. Use Object → Image Trace → Presets → "16 Colors" (or "High Fidelity Photo" for more detail).
3. Click Expand to convert to editable vector paths.
4. Clean up: simplify paths, adjust colors to exact brand hex codes, remove any AI artifacts.
5. This converts AI-generated illustrations to actual editable vector artwork.