LibraryMidjourney Omni-Reference Complete Guide: Master High-Fidelity Image Embedding
Midjourney's Omni-Reference is a groundbreaking V7 feature that lets you "put THIS in your image" by embedding characters, objects, vehicles, or creatures from any single reference image into your AI-generated artwork. Available on both the web UI and Discord, Omni-Reference is accessed via a drag-and-drop bin or the --oref <image_url> command, with influence controlled by the --ow (omni-weight) parameter ranging from 1 to 1,000 (default 100). Although it consumes double the GPU time of a standard V7 render, its precision makes it invaluable for creators seeking consistent, high-fidelity results.
What Is Midjourney Omni-Reference?
Omni-Reference is Midjourney's universal image-reference system introduced in V7, designed to embed any visual element from a reference image—people, props, vehicles, or non-human creatures—directly into generated images. Unlike the former V6 character references, it works with personalization, moodboards, and style references but isn't compatible with inpainting, outpainting, Draft Mode, or Fast Mode. Using Omni-Reference automatically doubles the GPU time required per render compared to a standard V7 job.
Key Features of Omni-Reference:
- Embed any visual element, from characters to objects
- Fully compatible with V7 model
- Adjustable weight parameter controls reference application strength
- Supports combination with personalization and style references
Limitations:
- Only one reference image allowed per prompt
- Consumes double the GPU time of standard V7
- Not compatible with inpainting, outpainting, Draft and Fast modes
- May trigger stricter content moderation
How to Use Omni-Reference
On the Web
- Switch your model to V7 in the Settings menu
- Click the image icon in the Imagine bar, then upload or select a reference image
- Drag that image into the Omni-Reference bin—only one image per prompt is allowed
- Adjust the omni-weight slider or append --ow <value> (1–1000, default 100) to control how strictly the reference is applied
On Discord
- Add --oref <image_url> to the end of your prompt with a valid online image URL
- Use --ow <value> to set the omni-weight; higher values (e.g., 400+) enforce stronger fidelity, while lower values (e.g., 25) allow more stylization
Best Practices and Tips
- Combine with text prompts: Always include clear descriptive text alongside your reference to convey scene details not present in the image
- Balance style and fidelity: Lower --ow (e.g., 25) when applying heavy style transfers (photo → anime); raise it (e.g., 400) to preserve details like facial features or clothing
- Multi-subject images: Use a single reference containing multiple characters or objects and explicitly mention each to have them all appear
- Moderation checks: Be aware that Omni-Reference may trigger stricter content moderation; blocked jobs cost no credits and only successful renders deduct GPU time
Weight Value Reference Table
| Weight Value | Effect | Recommended Uses |
|---|---|---|
| 1-25 | Very light reference influence, mostly follows prompt style | Heavy style transformations, creative interpretations |
| 50-100 | Balanced reference influence, preserves basic features but allows creativity | General purpose, default choice for most scenarios |
| 200-400 | Strong reference influence, high preservation of features and details | Preserving specific character facial features, brand elements |
| 500-1000 | Extremely strong reference influence, almost exact copying of reference elements | Professional applications requiring maximum fidelity |
Case Studies
Omni-Reference can be particularly effective in the following scenarios:
- Character Consistency: Maintaining the same character across multiple scenes and settings
- Product Visualization: Placing specific products in various contexts while preserving brand identity
- Style Transformations: Converting realistic references into stylized art while maintaining recognizability
- Complex Object Integration: Embedding detailed objects like vehicles or architecture into new environments
Limitations and Considerations
While Omni-Reference offers powerful capabilities, it's important to be aware of its limitations:
- Results may vary based on the complexity and clarity of the reference image
- Double GPU time consumption may impact your credit usage
- Some complex transformations might require several attempts to achieve optimal results
- The feature is still evolving and may be enhanced in future updates
Conclusion
Omni-Reference elevates Midjourney's creative toolkit by allowing precise embedding of any reference image element into your artwork. Master the --oref parameter, fine-tune the --ow omni-weight, and pair with detailed text prompts to achieve consistently high-fidelity, personalized images. Dive in, experiment freely, and share your Omni-Reference masterpieces with the community!