Artlist’s Image to Image generator allows you to blend up to 3 images. Clearer prompts equal cleaner, more accurate results. You can perform actions such as removing the glasses on a character, making a cloudy day sunny, and converting an image to a different perspective or style.
1. Be specific about people and objects
Instead of vague terms, clearly name and identify each element precisely who or what should be in the image.
Instead of: Everyone
Write: Man, woman, and child
2. Describe exactly what to change
AI reads your words literally. Spell out the exact alteration you want.
Instead of: Combine the images
Write: Put the toy as a graphic on the shirt
3. Use clear action verbs
Avoid ambiguous words. Tell the AI exactly what to do.
Instead of: Transform the scene Write: Replace the background
4. Edit text with quotation marks
“When replacing text in an image, put both the original and the new text in quotes so the AI can distinguish between instructions and design elements.
Example: In the poster, replace "For Sale" with "Not for sale".
5. State what should stay the same
If certain aspects must remain unchanged, say so. Write
Example: “Keep the same camera angle, position, and framing.”
6. Be specific about placement
Think of your prompt like stage directions. Placement words such as “to the left,” “in the background,” or “beside the woman”, help the AI arrange elements spatially and avoid surreal compositions.
Example:“Place the man on the left and the woman in the center.”
7. Start with aligned references
For seamless results, choose source images that share similar lighting, perspective, and tone. Mixing a bright photo with a soft sketch can create mismatched textures or depth. Keep your visual style cohesive from the start.
8. Control style drift with weighted prompts
If one image should dominate the blend, specify it directly.
Example: “Keep the woman from image 1 as the main subject. Blend background from image 2 softly.”
This maintains your key subject’s integrity while harmonizing supporting elements.
9. Think like a cinematographer
Guide the AI with cinematic cues. Directional, lighting, and mood descriptors help generate more grounded, atmospheric results.
Examples:
- “Shot at eye level”
- “Soft natural light, golden hour”
- “Shallow depth of field”
10. Build progressively
The AI works best with an iterative approach. Instead of overloading your first prompt, start simple and refine gradually.
Step 1: Blend the images cleanly.
Step 2: Adjust composition, lighting, or style in subsequent prompts.
This step-by-step process delivers smoother, more cinematic outcomes.
Explore more on AI-models
Improving prompts: Writing tips and Artlist's AI Enhancer
Prompting tips for Image to Image Models
AI Video models explained: choosing the right model for your project
Need some more help? Head back to our Help Center.