AI Beauty

Resources

How to integrate virtual makeup try-on on a brand website

From demo to catalog prep, widget/white-label deploy, and pilot metrics — without guaranteed timelines.

Integrating makeup virtual try-on is a product and operations project, not a one-line script install. This guide walks ecommerce and tech leads through a realistic path: demo → catalog slice → white-label or widget → pilot metrics. Timelines depend on SKU readiness and placement — we do not promise a universal “go live in N days.”

When try-on on the brand site is worth it

• Many near-neighbor shades (foundations, lips, blush, eyes) create purchase hesitation;
• returns or support tickets mention “wrong color,” not only logistics;
• you can show your SKUs — not a generic third-party palette;
• a named owner (ecommerce / digital / product) will agree pilot metrics.

If the ask is “AR for an investor deck,” a pilot rarely pays for the team’s attention. If the ask is reducing shade uncertainty on PDP and collecting shade signals, the format is justified.

Widget vs SDK vs API (neutral map)

• Widget / white-label page — fastest way to test try-on on a brand catalog with controlled branding.
• API — useful when you already own UI and need server-side or photo pipelines.
• SDK / Face AR engine — maximum control for apps and custom web; heavier integration and commercial process.

AI Beauty focuses on demo → partial-catalog pilot → white-label on the brand site. See Virtual Try-On and Pilot. A live white-label example is the Anaminerals case on podbor.anami.su.

We do not currently ship a public Shopify App Store listing as a primary distribution channel — if that is a hard requirement, validate it explicitly with any vendor (including us) before procurement.

Photo vs live camera UX

Photo try-on fits PDP and campaign landings: low friction, works when camera permission is denied. Live camera fits “mirror” moments and assisted selling. Many brands start with photo, then add camera once consent copy and mobile layout are stable.

SKU digitization basics

Pilot quality tracks catalog quality. Minimum fields usually include stable SKU id, display name, category, color reference (swatch/packshot), finish, PDP URL, and pilot inclusion flag. Start with one high-uncertainty category (foundations or lips) rather than the full assortment — do not digitize the entire range on day one.

PDP placement

• “Try on” on shade PDP;
• category landing with a curated shade set;
• campaign URL for paid traffic;
• post-try-on CTA back into cart, wishlist, or consult — keeping the shopper in your brand context.

Decide early whether try-on stays inside your funnel or exits to a third-party domain. White-label preserves brand context.

Privacy and consent

Spell out why a photo or camera is needed, keep a fallback when permission is denied, and align wording with your counsel. Try-on is shade visualization — not a medical skin diagnosis. Roadmap modules (for example Skin Advisor) should not be sold as live try-on features.

Pilot KPIs (no borrowed uplift)

• opens → photo/camera activation → completed try-on;
• shade switches and PDP clicks from the result;
• leads attributed to the pilot URL / sourcePage;
• qualitative mobile issues.

Comparing “site-wide conversion before/after” without traffic and assortment controls is a weak conclusion. Prefer funnel behavior inside the pilot perimeter. Vendor selection criteria: buyer’s guide.

Launch checklist

• Pilot SKU list approved by brand ops.
• Placement and post-try-on CTA agreed.
• Consent copy reviewed.
• Owner and review date set.
• Bug escalation channel for shade rendering and mobile UX.
• Success criteria written before traffic starts.

Request a pilot discussion with your category and storefront stack — we will scope a catalog slice without inventing conversion guarantees.

Request a pilot