Short recommendation
Use Leonardo AI’s photoreal model + fixed prompt template, fixed camera parameters, and either the built-in transparent-output/inpainting route or an automated background-removal step. Lock model + seed + lighting description and generate in 1:1 2048px or higher. Batch via Leonardo’s API or the studio batch feature and run a quick QA script for shadow/scale consistency before export to PNG (sRGB, transparency).

Why this works
Photoreal models produce crisp edges and realistic soft shadows when you explicitly describe light direction, contact shadow, camera lens, and background. Fixing seeds, viewpoint, and prompt template across runs is what makes a set consistent and Shopify-ready.

Quick reproducible workflow (step-by-step)
1) Template prompt (copy/paste and replace placeholders):
“Studio product shot of {PRODUCT_NAME}, photorealistic, extremely sharp details, centered on a white seamless background, single soft key light from top-left, soft fill from right, subtle top rim light, soft contact shadow directly beneath the product, 50mm lens, f/8, neutral white balance, accurate colors, no text, no other objects.”
Negative prompt: “watermark, logo, text, multiple items, colored background, blur, artifacts.”

2) Settings to start with
- Model: Leonardo photoreal / high-detail diffusion variant
- Aspect: 1:1 (square) for Shopify thumbnails; 2048x2048 (or 3000 px on longest side for zoom)
- Sampler: Euler a or DPM++ (default is fine)
- Steps: 20–35 (30 for higher detail)
- Guidance/CFG: 7–12 (raise if images drift from prompt)
- Seed: fix one per product or use sequential seeds for variations
- Variations: use small variation/modify for color or angle changes

3) Shadow consistency tips
- Add explicit phrasing: “soft contact shadow, 45° shadow fall, shadow opacity 10–20%” and “no cast shadows from other objects.”
- Use the same camera angle text (e.g., “eye-level, 20° top-down, centered”) and same seed to keep shadow position consistent.
- If Leonardo supports an alpha output, request “transparent background” so shadows remain separate layers; otherwise render on white and remove background with inpainting/mask.

4) Bulk generation & automation
- Create a CSV with columns: product_name, short_desc, seed, color_override. Script merges into the prompt and calls Leonardo API or Studio batch.
- For many SKUs: generate 3–5 variations per SKU, pick the best using a simple color/edge QA script, then re-run refined prompts if needed.
- Use an automation step to convert outputs to PNG, ensure sRGB, and strip EXIF.

Decision criteria (which method to use)
- Use in-studio transparent/alpha outputs if you can: fastest, easiest, best for layering.
- Use inpainting to replace background with perfect white if alpha isn’t available.
- If you have many SKUs and coding resources: use API + CSV. If one-off or low volume: Studio manual batches are fine.

Best-for / Avoid-if
- Best for: e‑commerce teams needing consistent catalog images, small studios without full photographic setups.
- Avoid if: you need pixel-perfect color matching for high-end fashion (use a pro photo pipeline or real photo plus AI cleanup).

Practical QA checklist before upload
- Colors checked vs. reference (sample 5%)
- Shadows aligned and similar density across set
- PNG exported with transparency or pure white background
- sRGB color profile applied
- Size/crop centered with ~10% padding
- Filenames SKU-friendly and consistent

If you want, I can give a ready-to-run CSV + prompt template and a small Node/Python snippet to call Leonardo’s API and produce PNGs automatically.

Compare Leonardo AI and Midjourney