AI Image Generation

Create photorealistic interior visualizations using AI to show materials in context and present design concepts to clients.

7 min read

Transform your material selections into photorealistic interior visualizations using Palett's AI image generation. Perfect for client presentations, design exploration, and showing materials in realistic contexts.

What is AI Image Generation?

AI image generation creates:

  • Photorealistic room concepts from text descriptions
  • Material visualizations showing your selections in context
  • Design mockups for client presentations
  • Iterative variations to explore different directions
  • High-quality images suitable for presentations and portfolios

Powered by: Advanced generative AI models trained on interior design imagery

Getting Started

Access AI Generation

Three ways to generate:

  1. Visualize Tab: Dedicated workspace for AI sessions

    • Go to project → Visualize
    • Click "New Session" or "+ Generate"

  2. AI Chat: Conversational generation

    • Open chat → "Generate a modern living room concept"
    • AI guides you through the process

  3. Quick Generate: From boards or materials

    • Select materials → "Visualize in Room"

Your First Generation

Simple 3-step process:

  1. Describe your vision:

    "Modern living room with natural light, mid-century furniture, 
 and neutral color palette"

  2. (Optional) Attach references: Add up to 14 images from your boards

  3. Click "Generate": Wait 30-60 seconds

Writing Effective Prompts

Prompt Structure

Good prompt includes:

  1. Room type: Living room, kitchen, office, hotel lobby
  2. Style: Modern, traditional, coastal, industrial
  3. Key features: Natural light, high ceilings, open concept
  4. Mood/atmosphere: Warm and inviting, minimalist, luxurious
  5. Specific elements: Fireplace, built-ins, large windows

Example prompt:

"Spacious modern kitchen with waterfall island, brass fixtures, 
 white marble backsplash, natural oak cabinets, and abundant 
 natural light through large windows. Warm and inviting atmosphere."

Prompt Best Practices

Be specific but not overly prescriptive:

✅ Good: "Modern living room with a large sectional sofa and abstract art"

❌ Too vague: "Nice living room"

❌ Too prescriptive: "Living room with sofa at 45-degree angle, exactly 3 throw pillows, lamp 2 feet to the left"

Describe atmosphere, not just objects:

✅ Good: "Cozy, warm, inviting space with soft textures"

❌ Object-focused: "Room with couch, table, lamp, rug"

Include lighting cues:

  • "Natural light streaming through windows"
  • "Warm ambient lighting"
  • "Dramatic spotlighting on artwork"
  • "Soft diffused light"

Mention scale when relevant:

  • "Intimate breakfast nook"
  • "Grand hotel lobby with 20-foot ceilings"
  • "Compact urban studio apartment"

Using Reference Images

Adding References

Attach up to 14 reference images to guide the AI:

From your boards:

  1. Click "Add References"
  2. Select materials from your project boards
  3. AI incorporates those materials into the design

Upload inspiration:

  1. Click "Upload"
  2. Add client mood board images, Pinterest photos, style references
  3. AI matches the aesthetic

Types of References

Material references (from your boards):

  • Flooring samples
  • Wall finishes
  • Fabric swatches
  • Tile patterns
  • Lighting fixtures

Style references (inspiration images):

  • Pinterest room photos
  • Magazine clippings
  • Client's favorite spaces
  • Competitor designs (for differentiation)

Spatial references:

  • The actual room to be designed (before photos)
  • Similar room layouts
  • Architectural details

Reference Strategy

For material-focused visualizations:

Attach: 5-8 specific materials from your boards
Prompt: "Modern bathroom featuring these materials"
Result: AI creates scene showcasing your selections

For style exploration:

Attach: 3-5 inspiration images with desired aesthetic
Prompt: "Coastal living room in this style"
Result: AI interprets the style, generates similar vibe

For client presentations:

Attach: Client's existing space photo + your material selections
Prompt: "Redesign this space with these materials"
Result: Shows transformation with your specified materials

Visualization Modes

Choose how AI interprets your prompt and references:

Room Scene (Default)

What it creates: Photorealistic interior with materials in context

Best for:

  • Client presentations
  • Showing materials in realistic settings
  • Portfolio pieces

Example:

Prompt: "Modern kitchen with the attached marble and brass fixtures"
Result: Full kitchen scene with your materials integrated naturally

Mood Board

What it creates: Styled arrangement showing all materials together

Best for:

  • Material palette presentations
  • Quick client review of selections
  • Comparing multiple material combinations

Example:

Prompt: "Create mood board with these materials"
Result: Artistic layout showing materials as cohesive collection

Material Study

What it creates: Close-up details and texture focus

Best for:

  • Highlighting material quality
  • Showing texture and detail
  • Technical presentations

Example:

Prompt: "Close-up study of this tile pattern"
Result: Detailed view showing tile texture, grout, finish

Inspiration

What it creates: Creative interpretation, artistic freedom

Best for:

  • Exploring new directions
  • Early concept phase
  • Creative brainstorming

Example:

Prompt: "Interpret these materials in an unexpected way"
Result: AI takes creative liberties, suggests novel applications

Generation Settings

Quality Settings

Draft (2 credits):

  • Faster generation (~30 seconds)
  • Lower resolution (1024x1024)
  • Good for internal review and iteration

Standard (4 credits):

  • Balanced speed and quality (~45 seconds)
  • Medium resolution (1536x1536)
  • Good for most client presentations

High Quality (6 credits):

  • Slower generation (~60 seconds)
  • High resolution (2048x2048)
  • Best for final presentations and prints

Aspect Ratio

  • Square (1:1): Balanced, versatile
  • Landscape (16:9): Wide rooms, panoramic views
  • Portrait (9:16): Vertical spaces, tall ceilings
  • Custom: Specify exact dimensions

Iterating and Refining

Generating Variations

Not happy with the first result? Refine it:

Adjust the prompt:

Original: "Modern living room"
Refined: "Modern living room with more natural light and lighter colors"

Add/remove references: Change which materials or images are attached

Change mode: Try "Material Study" instead of "Room Scene"

Adjust settings: Increase quality, change aspect ratio

Iteration Strategy

Start broad, then narrow:

  1. Round 1: General concept

    • "Modern office space"
    • Generate 2-3 options

  2. Round 2: Refine preferred direction

    • "Modern office with natural wood and green accents"
    • Generate 2-3 variations

  3. Round 3: Final polish

    • "Modern office with oak desk, sage walls, brass lighting"
    • Generate 1-2 finals

Cost: 3 rounds × 3 images × 4 credits = 36 credits

Common Refinements

Lighting adjustments:

  • "Brighter" / "More dramatic lighting"
  • "Add natural light from windows"
  • "Warmer lighting"

Color tweaks:

  • "Lighter palette"
  • "More colorful"
  • "Neutral tones only"

Composition changes:

  • "Different camera angle"
  • "Wider view"
  • "Focus on the fireplace area"

Best Practices

1. Generate Multiple Options

Don't settle for first result:

  • Generate 3-5 initial concepts
  • Show client variety
  • Identify which direction resonates

Budget: 3-5 images × 4 credits = 12-20 credits

2. Use References Strategically

Too few references (0-2): AI has creative freedom but may not match your vision

Right amount (5-8): AI has guidance but room for creativity

Too many (12-14): Can confuse AI, conflicting styles

Sweet spot: 5-8 references that tell a cohesive story

3. Start with Draft Quality

Workflow:

  1. Generate 5 draft concepts (2 credits each = 10 credits)
  2. Client picks favorite
  3. Regenerate that one in high quality (6 credits)

Savings: 16 credits vs. 30 credits (if all were high quality)

4. Save Everything

Even "failed" generations:

  • Document what didn't work
  • Learn what prompts/references work best
  • Show client evolution of design

5. Organize in Sessions

Create folders in Visualize tab:

  • "Option A - Modern"
  • "Option B - Traditional"
  • "Client Rejected"
  • "Finals"

Keeps iterations organized.

Advanced Techniques

Material Swapping

Scenario: Client approved design but wants different material

Method:

  1. Open approved visualization
  2. Click "Remix"
  3. Remove old material reference, add new material
  4. Keep same prompt
  5. Generate

Result: Same room, different material

Style Transfer

Scenario: Client loves Room A's layout but Room B's style

Method:

  1. Use Room A photo as reference
  2. Prompt: "Redesign in the style of [Room B style]"
  3. Generate

Result: Room A layout with Room B aesthetic

Before/After Comparisons

Create client-ready presentations:

  1. Generate "before" (current space)
  2. Generate "after" (your design)
  3. Place side-by-side in PDF export
  4. Show transformation

Troubleshooting

"Result doesn't match my materials"

Causes:

  • Too many conflicting references
  • Prompt overrides material references
  • Materials not clear in reference images

Solutions:

  • Reduce to 5-8 key materials
  • Be specific: "Feature the attached marble tile prominently"
  • Use clearer material photos (close-ups, good lighting)

"AI ignored my prompt"

Causes:

  • Prompt conflicts with references
  • Ambiguous language
  • Asking for impossible combinations

Solutions:

  • Simplify prompt, be more direct
  • Remove conflicting references
  • Try different phrasing

"Images look unrealistic or weird"

Causes:

  • Unusual or rare combinations
  • Conflicting style references
  • AI hallucination

Solutions:

  • Regenerate (results vary)
  • Adjust references to be more coherent
  • Try different mode (Material Study instead of Room Scene)

Credit Management

Costs per image:

  • Draft: 2 credits
  • Standard: 4 credits
  • High Quality: 6 credits

Monthly allocations:

  • Free: 25 one-time starter credits
  • Professional: 350 credits/month
  • Studio: 800 credits/month
  • Agency: 2,500 credits/month

Strategies:

  • Use drafts for iteration (save 2-4 credits per image)
  • Generate high-quality only for finals
  • Buy credit packs if you exceed monthly allocation

See Understanding Credits for details.

What's Next?

Need help? Chat with us (bottom right) or email support@palett.space.

Was this page helpful?