How to Use Flat-to-Model: Turn Product Photos into On-Model Images

Every fashion brand needs on-model photography. But booking studios, hiring models, and coordinating photoshoots for every SKU in your catalog is slow, expensive, and doesn't scale. Flat-to-Model changes that.

With On-Model's Flat-to-Model feature, you upload flat-lay product photos — the kind you already have — and get back professional-quality on-model images in minutes. No photoshoot required.

This guide walks you through the entire process in the On-Model app, from uploading your first garment to downloading the final result.

What you'll need

• An On-Model account — sign up free if you haven't already
• Flat-lay product photos of the garments you want to transform (top, bottom, and/or footwear)
• A few minutes

Step 1: Upload your product images

The Flat-to-Model wizard starts with the upload step. You'll see three slots for your outfit:

• Top — Shirts, jackets, sweaters, blouses
• Bottom — Trousers, skirts, shorts
• Footwear — Shoes, boots, sandals

Drag and drop or click to upload an image into each slot. You can upload just one item (like a top only) or a complete outfit with all three pieces. Here's the full outfit we'll use in this guide:

Step 2: Select your AI model identity

Next, you'll choose the AI model identity that will "wear" your garments. The identity selector shows a grid of available model faces.

You can filter by:

• Default — On-Model's curated library of diverse models
• Mine — Custom identities you've created
• Shared — Identities shared within your team or organization

Pick the model that best matches your brand's target audience. Each identity produces consistent results across all your products, so you can maintain a uniform look across your entire catalog. For this guide, we're using Elise:

Step 3: Configure your instructions

This is where you tell On-Model exactly how you want the final image to look. You have two options:

Quick Start presets

Choose from pre-built presets designed for common use cases:

• PDP — Clean product detail page shots (white background, front-facing pose)
• Lifestyle — Natural, lifestyle-oriented shots with contextual backgrounds
• Social — Optimized for social media content (dynamic poses, bold compositions)
• Editorial — Magazine-style editorial photography

Custom instructions

Or build your own instructions from scratch:

• Pose — Standing, walking, sitting, dynamic, or describe any specific pose
• Background — White studio, urban street, indoor setting, or any scene
• Style — Photography style (editorial, commercial, candid, etc.)
• Camera — Angle, lens, and framing preferences
• Lighting — Studio, natural, dramatic, soft, etc.

You'll also set output options:

• Size — 1K, 2K, or 4K resolution
• Aspect ratio — 1:1, 3:4, 4:3, 9:16, or 16:9
• Format — PNG or JPG
• Variations — Generate 1-4 variations per instruction

For this guide we used: standing front-facing pose, white studio background, 2K resolution, 3:4 aspect ratio.

Step 4: Review and start processing

The final step shows a summary of everything you've configured:

• SKU images — The garments you uploaded
• Identity — The AI model selected
• Instructions — Your pose, background, and style settings
• Total outputs — How many images will be generated
• Credit cost — How many credits this job will use

Give your project a name (this helps you find it later in the jobs list), then hit Start Processing.

You'll be redirected to the Jobs page where you can track your job's progress in real time. Processing typically takes 1-3 minutes depending on the number of outputs and current queue.

Your results

Once processing completes, you'll find your finished on-model images in the job results. Here's the full transformation — from three flat-lay inputs to a professional on-model photograph:

The AI combines all three garment pieces into a single cohesive outfit, preserving the exact appearance of each item — colors, textures, patterns, and proportions — while generating a natural, realistic pose on the model you selected.

Tips for best results

1. Use high-resolution images — Higher quality inputs produce higher quality outputs. Aim for at least 1000px on the longest side.

2. Clean backgrounds matter — Flat-lay photos on white or light neutral backgrounds work best. Busy backgrounds can confuse the garment detection.

3. Upload complete outfits — The more garment pieces you provide, the more realistic the final result. A polo alone looks good; a polo + trousers + shoes looks great.

4. Experiment with presets — Try the same input with different presets (PDP vs. Lifestyle vs. Social) to see which style works best for your use case.

5. Use consistent identities — Pick one or two AI model identities and use them across your entire catalog for a cohesive brand look.

6. Try multiple variations — Set variations to 2-3 to get different poses and compositions from the same input, then pick your favorite.

What's next

You've just created your first on-model image from a flat-lay photo. From here, you can:

• Scale up — Submit batch jobs with hundreds of products at once
• Try Model Swap — Already have on-model photos? Swap the model while keeping the garment pixel-perfect
• Integrate via API — Connect On-Model directly to your product catalog pipeline

Ready to start? Head to the On-Model app and try it with your own products.