> ## Documentation Index
> Fetch the complete documentation index at: https://withlooksy.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# How Accurate is Virtual Try-On?

> Looksy's AI virtual try-on delivers realistic previews of style, color, and fit. Learn what to expect and how accuracy improves with quality photos.

## Virtual Try-On Accuracy Explained

Looksy's AI-powered virtual try-on delivers **realistic visual previews** that help shoppers see how garments will look on them. Accuracy depends on several factors, including photo quality, garment type, and the technology used.

## What Virtual Try-On Does Well

### Highly Accurate Areas

<CardGroup cols={2}>
  <Card title="Style & Silhouette" icon="shirt">
    **95%+ accuracy**

    The overall shape, cut, and silhouette of the garment are rendered accurately across different body types.
  </Card>

  <Card title="Color & Pattern" icon="palette">
    **90%+ accuracy**

    Colors, graphics, and patterns are displayed realistically. Customers can see how prints and designs look on them.
  </Card>

  <Card title="Placement & Positioning" icon="crosshairs">
    **90%+ accuracy**

    Garments are positioned correctly on the body, showing necklines, sleeves, and hem placement accurately.
  </Card>

  <Card title="Overall Visualization" icon="eye">
    **85%+ accuracy**

    The combined effect gives shoppers a strong sense of how the item will look on them, reducing purchase uncertainty.
  </Card>
</CardGroup>

## What Virtual Try-On Approximates

### Areas with Limitations

<AccordionGroup>
  <Accordion title="Fabric Texture & Drape">
    **Moderate accuracy**

    The AI can show general drape and flow, but fine texture details (like cable knit patterns or bouclé fabric) may not render perfectly.

    **What this means:** Shoppers see the overall look, but very specific fabric feel is approximated.
  </Accordion>

  <Accordion title="Precise Fit & Sizing">
    **Not a substitute for size charts**

    Virtual try-on shows how the style looks, but it's **not a sizing tool**. It doesn't predict whether a Medium or Large will fit better.

    **What this means:** Always pair virtual try-on with detailed size charts and fit guides.
  </Accordion>

  <Accordion title="Complex Embellishments">
    **Variable accuracy**

    Heavy sequins, 3D appliques, and intricate beading may not render with full detail.

    **What this means:** The overall look is accurate, but very fine details may be simplified.
  </Accordion>

  <Accordion title="Sheer or Transparent Fabrics">
    **Moderate accuracy**

    Sheer materials, mesh panels, and see-through fabrics can be challenging to render realistically.

    **What this means:** Test these products first to ensure the result meets your quality standards.
  </Accordion>
</AccordionGroup>

## Factors That Improve Accuracy

### Product Image Quality

**What actually helps accuracy:**

<Check>
  ✓ **Multiple angles** (max 4 images per product) → Better AI context

  ✓ **Detail shots** (close-ups of texture/material) → More realistic rendering

  ✓ **Any background works** (plain, complex, or busy) → AI handles all backgrounds

  ✓ **Higher resolution helps** (1200px+ recommended) → Clearer details

  ✓ **Showcasing garment details** (unique features) → More accurate results
</Check>

### Selfie Quality

**What helps with selfies:**

* **Full body shot** – Shows complete garment placement (works best)
* **One person only** – Required for AI to process correctly
* **Human photo** – No cartoons, animals, or illustrations
* **Enough body visible** – At minimum, enough to show the garment

<Note>
  Don't worry about expression, sunglasses, hats, facing camera directly, or recent photos. The app works with any selfie.
</Note>

### Garment Type

**Some products work better than others:**

* **Best:** Tops, t-shirts, blouses, dresses with clear structure
* **Good:** Jackets, hoodies, activewear
* **Variable:** Oversized styles, drapey cardigans, highly textured fabrics

## How Looksy Compares to Competitors

### Industry-Leading Accuracy

Looksy uses state-of-the-art AI models trained on millions of garments and body types, delivering:

* **Faster rendering** – Results in 15-20 seconds
* **More realistic outputs** – Advanced fabric and shadow rendering
* **Better body type diversity** – Trained on diverse datasets for inclusive representation
* **Consistent quality** – Reliable results across product categories

### Continuous Improvement

The AI models are constantly updated with:

* New garment types and styles
* Improved fabric rendering
* Better handling of complex patterns
* Enhanced accuracy for diverse body types

## Realistic Customer Expectations

### What to Tell Shoppers

**Do say:**

* "See how the style and color look on you"
* "Visualize the garment before you buy"
* "Get a realistic preview of the product"

**Don't say:**

* "Perfectly accurate representation"
* "Exact fit prediction"
* "100% photorealistic rendering"

### Combine with Other Tools

Virtual try-on is most effective when combined with:

* **Detailed size charts** – For fit guidance
* **Fit descriptions** – "Runs small," "oversized fit," etc.
* **Customer reviews** – Real feedback on fit and quality
* **Product videos** – Show fabric movement and texture

<Note>
  Virtual try-on is a visualization tool, not a sizing tool. It reduces uncertainty by showing style and appearance, but doesn't replace size charts.
</Note>

## Measuring Accuracy Impact

### How Merchants Can Assess Performance

Track these metrics to see if accuracy meets customer needs:

1. **Return rate** – Are returns lower for try-on products?
2. **Conversion rate** – Do products with try-on convert better?
3. **Customer feedback** – What are shoppers saying about the feature?
4. **Engagement rate** – Are customers using try-on repeatedly?

If these metrics improve, the accuracy is meeting customer expectations.

### What Good Performance Looks Like

Stores using Looksy typically see:

* **10-25% conversion rate increase** on try-on enabled products
* **15-30% reduction in return rates** due to better expectations
* **2-3x higher engagement** on product pages with try-on
* **Positive customer feedback** on the feature's helpfulness

## Common Accuracy Questions

<AccordionGroup>
  <Accordion title="Will it show exactly how the garment fits me?">
    Virtual try-on shows how the style looks on you, but it's not a precise fit predictor. It helps you visualize appearance, but you should still refer to size charts for fit guidance.
  </Accordion>

  <Accordion title="Why does the result sometimes look slightly off?">
    Accuracy depends on photo quality. Poor lighting, angled selfies, or low-resolution product images can reduce accuracy. Using high-quality photos improves results significantly.
  </Accordion>

  <Accordion title="Can it handle all body types?">
    Yes. Looksy is trained on diverse body types and works across different shapes, sizes, and skin tones. The AI adapts the garment rendering to match the person in the selfie.
  </Accordion>

  <Accordion title="Does accuracy improve over time?">
    Yes. The AI models are continuously updated with new data, improving rendering quality, fabric accuracy, and support for more garment types.
  </Accordion>
</AccordionGroup>

## Transparency About Limitations

### We're Honest About What Virtual Try-On Can't Do

**Virtual try-on is not:**

* A replacement for trying on clothes in person
* A precise fit or sizing tool
* A perfect fabric texture simulator
* A substitute for customer reviews and size charts

**Virtual try-on is:**

* A powerful visualization tool
* A way to reduce purchase uncertainty
* A method to see style and color on yourself
* A tool to increase confidence and reduce returns

<Check>
  By setting realistic expectations, you'll have happier customers who understand what virtual try-on offers and use it effectively.
</Check>

## Improving Accuracy for Your Store

### Actionable Steps

1. **Audit product images** – Ensure they meet quality standards
2. **Test across product types** – Identify which categories work best
3. **Gather customer feedback** – Ask what shoppers think of the accuracy
4. **Iterate and improve** – Disable underperforming products, expand winners
5. **Educate customers** – Set proper expectations about what try-on shows

## Next Steps

<CardGroup cols={2}>
  <Card title="Image Requirements" icon="image" href="/product-setup/image-requirements">
    Optimize product photos for maximum accuracy
  </Card>

  <Card title="Best Practices" icon="star" href="/product-setup/best-practices">
    Strategies to maximize try-on performance
  </Card>
</CardGroup>