Uni-1 is our first unified model designed for both image creation and precision editing. What makes it powerful isn’t just output quality—it’s control. You can generate entirely new images, or surgically modify existing ones, all while guiding the system with references, seeds, and structured prompts.
This guide gives you a working mental model and practical workflows—from your very first image to advanced multi-reference setups—so you can move from exploration to repeatable results.
What makes Uni-1 different
Uni-1 stands out because it combines flexibility with structure:
- Two clear modes
- Create Image → generate something new
- Modify Image → edit something existing
- Up to 9 reference images, each with a defined role
- Strong control over niche and specific visual styles
- Seed support for reproducibility and controlled iteration
- Nine aspect ratios, from ultra-tall to ultra-wide
If you understand the modes and how to control references, you understand Uni-1.
The core distinction: Create vs Modify
Everything in Uni-1 starts with a simple question:
Am I creating something new, or changing something that already exists?
Create Image
- Produces a brand-new composition
- Can be inspired by references, but does not preserve them
Modify Image
- Edits a specific input image
- Preserves composition and structure unless told otherwise
Examples:
Modify:
Make this photo look like nighttime
Create:
Create a new scene in the style of this photo
When in doubt:
- If the output should look like a version of your input, use Modify
- If it should feel inspired but new, use Create
Your first image
Start simple:
- Open Create Image
- Write a prompt
- Choose an aspect ratio
- Generate
Example prompt:
A wide cinematic landscape at golden hour. Rolling green hills, a lone tree in the foreground, warm orange light, dramatic clouds. Photorealistic.
Don’t overthink it. The real workflow is:
prompt → evaluate → refine
Leave the seed blank while exploring. Once you find something strong, lock it and iterate from there.
Understanding the core parameters
Prompt
Your primary control. Up to 6,000 characters. Be precise.
Aspect ratio
Controls framing, not quality. Choose based on use case.
Seed
Controls randomness:
- Same seed + same prompt → same result
- Same seed + changed prompt → controlled variation
- No seed → exploration
Reference images (Create)
Up to 9 images to guide different aspects.
Source image (Modify)
The image you are editing. Dimensions are preserved automatically.
Working with reference images
References only work if you tell the model what they are for.
Use this structure:
Use IMAGE1 ([description]) as a [ROLE] reference.
Possible roles:
- Style
- Character
- Composition
- Color palette
- Lighting
- Texture
- Mood
Without roles, the model guesses—and guesses are unreliable.
Example prompts
Create mode
Style reference
Use IMAGE1 (impressionist oil painting with loose brushwork and warm sunset tones) as a STYLE reference. Create a portrait of a woman in her 30s applying IMAGE1’s color palette and painterly texture. The subject should be new.
Character reference
Use IMAGE1 (woman with short copper-red hair, freckles) as a CHARACTER reference. Preserve her features. Generate a new scene: she’s sitting in a softly lit café.
Multi-reference
Use IMAGE1 as a COLOR PALETTE reference, IMAGE2 as LIGHTING, IMAGE3 as COMPOSITION. Create a lone figure walking through a rain-slicked street at night.
Modify mode
In Modify mode, clarity is everything.
Change the time of day to golden hour. Update sky, light direction, shadows, and color temperature. Keep all subjects and composition unchanged.
Make the rain heavier and add reflections on the ground. Keep all other elements unchanged.
Always specify:
- What to change
- What must stay untouched
Prompting guidelines
Recommended lengths:
- Text-to-image → 80–250 words
- Reference-guided → 100–300 words
- Modify → 30–100 words
Avoid:
- Vague terms (“beautiful,” “amazing”)
- Redundant phrasing
- Conflicting instructions
Do this instead:
Golden hour, 85mm lens, shallow depth of field
Uni-1 performs best with specific, named aesthetics:
1970s Italian giallo film poster, high-contrast color blocking
Precision beats generality.
Aspect ratios
Choose based on where the image will live:
- 1:1 → social posts
- 9:16 → vertical video
- 16:9 → widescreen
- 3:2 / 2:3 → photography
- 2:1 / 3:1 → cinematic/panoramic
- 1:2 / 1:3 → ultra-tall
In Modify mode, aspect ratio is locked to the source image.
Seeds: control and reproducibility
Seeds turn experimentation into systems.
- Fixed seed → consistency
- No seed → exploration
Workflow:
- Explore (no seed)
- Find a strong result
- Lock the seed
- Change one variable at a time
Save your prompt + seed together. That’s your reusable recipe.
Advanced techniques
Character consistency
- Generate a clean, front-facing reference image
- Reuse it as IMAGE1 (CHARACTER) in every scene
- Keep the label identical across prompts
Multi-reference architecture
Assign one role per image:
- IMAGE1 → character
- IMAGE2 → style
- IMAGE3 → lighting
- IMAGE4 → environment
End with:
Treat each reference as having authority over its assigned layer only.
Create → Modify chain
- Use Create to explore compositions
- Use Modify to refine details
This is one of the most powerful workflows.
Iterative refinement
- Explore (no seed)
- Lock seed
- Change one variable per generation
- Document results
It feels slower, but it’s actually faster because you always know what changed.
Troubleshooting
- References ignored → label each one clearly
- Modify changes too much → explicitly state what must stay unchanged
- Inconsistent outputs → lock the seed
- Prompt partially ignored → remove conflicts or split into steps
- Looks like the reference image → you may be in Modify mode
- Character inconsistency → reuse a canonical reference image
Quick reference
Create Image
- New compositions
- Text + up to 9 references
- Descriptive prompts
Modify Image
- Edits existing images
- Source image + references
- Direct, surgical prompts
The golden rules
- Label every reference
- In Modify mode, always state what should NOT change
- Change one variable at a time when refining
- Save prompt + seed for reproducibility
- Create = new scenes, Modify = edits
A helpful reminder
Uni-1 isn’t about getting the perfect result on the first try. It’s about building control over outcomes. The more structured your approach — clear roles, consistent seeds, deliberate iteration — the more predictable and powerful your results become.
Key takeaway
Mastering Uni-1 comes down to three things:
Choose the right mode → control your references → iterate with intention.
Once those are in place, you’re no longer guessing, you’re directing.
Uni-1 is not about writing the perfect prompt.
It’s about:
Start → Direct → Refine → Finish
You can prompt with:
- short text
- long detailed prompts
- structured formats (JSON)
- reference images
- sketches / layouts
1. The “Fast Start” Prompt (best for 80% of use cases)
Template
A [subject], in [style], with [lighting], [camera/composition], [environment/background], mood: [emotion], details: [key specifics]
Example
A ceramic artist shaping a lopsided bowl, documentary photography style, soft window lighting, close-up shot, cluttered home studio background, mood: focused and quiet, details: clay-covered hands, imperfect texture, tools scattered on wooden table
When to use
- Exploration
- First ideas
- Quick outputs
2. The “Cinematic Control” Prompt
Template
Subject: [who/what]
Style: [editorial / documentary / fine art / etc.]
Scene:
- Environment: [where]
- Time of day: [lighting context]
- Mood: [emotional tone]
Camera:
- Shot type: [close-up / wide / overhead]
- Lens: [35mm / 85mm / etc.]
- Composition: [framing approach]
Details:
- Textures: [materials]
- Motion: [if any]
- Extra: [small details that matter]
Example
Subject: A street tailor repairing a jacket
Style: documentary photography
Scene:
- Environment: sidewalk workshop with hanging fabrics
- Time of day: late afternoon natural light
- Mood: patient, intimate
Camera:
- Shot type: medium close-up
- Lens: 50mm
- Composition: slightly off-center, hands in focus
Details:
- Textures: worn fabric, frayed edges, thread spools
- Motion: subtle hand movement mid-stitch
- Extra: pins in mouth, chalk markings on fabric
When to use
- Editorial visuals
- Story-driven imagery
- High-quality outputs
3. The “Direct Edit” Prompt
Template
Edit instructions:
- Change: [what to modify]
- Keep: [what must stay the same]
- Style shift: [optional]
- Lighting: [optional]
- Details: [specific adjustments]
Example
Edit instructions:
- Change: replace background with a quiet bookstore interior
- Keep: subject posture, clothing, and facial expression
- Style shift: slightly more warm and nostalgic
- Lighting: soft golden indoor lighting
- Details: add bookshelves with uneven stacks and handwritten notes
Pro tip
Always define what to keep (this prevents drift)
4. The “Multi-Reference Fusion” Prompt
Template
Combine the following:
IMAGE 1: [what it provides]
IMAGE 2: [what it provides]
IMAGE 3: [optional]
Output:
- Subject: [final image goal]
- Style: [dominant style]
- Composition: [layout guidance]
- Details: [specific merges]
Example
Combine the following:
IMAGE 1: a vintage portrait (use for pose and framing)
IMAGE 2: a modern fashion editorial (use for styling and lighting)
IMAGE 3: textured paper background (use for surface quality)
Output:
- Subject: a contemporary portrait with vintage posture
- Style: editorial with subtle retro influence
- Composition: centered subject, clean framing
- Details: blend modern clothing with aged texture finish
When to use
- Style blending
- Character consistency
- Editorial work
5. The “Layout Control” Prompt
Template
Layout instructions:
[LEFT]: [object + description]
[CENTER]: [object + description]
[RIGHT]: [object + description]
[BACKGROUND]: [environment]
Style: [style]
Lighting: [lighting]
Example
Layout instructions:
LEFT: a stack of handwritten letters tied with string
CENTER: a wooden desk with an open notebook
RIGHT: a cup of tea with steam rising
BACKGROUND: softly lit room with a window
Style: still life photography
Lighting: soft morning natural light
Pro tip
Treat this like directing a photoshoot
6. The “Storyboard Generator”
Template
Create a storyboard with [X] frames.
Story:
[short narrative]
Frame details:
1. [scene]
2. [scene]
3. [scene]
Style: [style]
Consistency: same characters and setting
Example
Create a storyboard with 4 frames.
Story:
A baker wakes up early to prepare bread for the day.
Frame details:
1. Dim kitchen, flour being poured
2. Hands kneading dough on wooden surface
3. Oven light glowing with rising bread
4. Morning customers entering small bakery
Style: warm, documentary realism
Consistency: same space and character throughout
7. The “Loose / Creative Mode”
Template
[vibes, fragments, ideas, emotions]
Example
quiet morning, dust in sunlight, old apartment, slightly messy, books everywhere, slow life, soft textures, calm, lived-in feeling
When to use
- Mood exploration
- Unexpected ideas
- Early ideation
8. The “Structured JSON” Prompt
Template
{
"subject": "",
"style": "",
"environment": "",
"lighting": "",
"camera": {
"shot": "",
"lens": "",
"composition": ""
},
"details": []
}
Example
{
"subject": "elderly man repairing a watch",
"style": "fine art photography",
"environment": "small workshop filled with tools",
"lighting": "soft directional window light",
"camera": {
"shot": "close-up",
"lens": "85mm",
"composition": "tight framing on hands"
},
"details": ["tiny screws", "dust particles", "aged skin texture"]
}
Core Workflows
Workflow 1: Idea → Final Image
- Start with Fast Prompt
- Generate
- Refine with direct instructions
- Reroll
- Finalize
Workflow 2: Reference-Driven Creation
- Add 2–3 references
- Use fusion prompt
- Adjust composition
- Refine
Workflow 3: Fix & Polish
- Generate
- Identify issues
- Edit specific regions
- Repeat
Workflow 4: Precision Composition
- Generate base
- Add layout
- Label clearly
- Regenerate
Workflow 5: Exploration → Lock → Iterate
- Generate variations
- Pick favorite
- Lock seed
- Iterate small changes
Quick Rules
- Be specific (subject + style + composition)
- Put key instructions early
- Use references for consistency
- Don’t restart — refine
- Use layout for control
- Use editing for precision
- Iterate in small steps
A helpful reminder
You don’t need a perfect prompt.
You need:
a clear idea + direction + iteration
Uni-1 isn’t about prompting better.
It’s about:
thinking visually, then directing the system