studioGPT
Wear it before you buy it.
Shoot it without a studio.
One AI engine, two surfaces. Real-time virtual try-on on the product page —
and bulk on-model studio photography for the rest of your catalog.
Apparel-aware · Multi-tenant from day one · MVP demo available
MVP demo available — book a private walkthrough to see it in action.
Two jobs. One model. Zero duplicate spend.
Every other VTO vendor makes you choose: a consumer widget on the PDP, or a separate photo-generation tool for your catalog team. studioGPT runs both off the same AI backbone — and bills you as one premium add-on, not two.
Real-time try-on at the PDP
A "Try it on" button on every eligible apparel product. Shoppers pick or build an avatar, see themselves in the piece, and add to cart with confidence. Sits inside shopperGPT — no extra integration, no separate vendor.
- Pre-built diverse avatar library + custom avatars from a single selfie
- Multi-piece try-on: tops, bottoms, dresses, jumpsuits, shoes
- Privacy-first: person images are session-scoped and never trained on
Bulk on-model studio photography
Send a flat lay or ghost mannequin. Get a full diverse on-model shoot in return — across body types, poses, and demographics. Generated images flow straight into your product media, picked up by catalogGPT, and become the new default hero image set.
- Replaces $300–$1,500-per-SKU studio photography line items
- Idempotent per SKU × avatar × seed — re-runs do not duplicate or corrupt
- Cost-controlled per merchant: monthly budget + hard cap + dry-run estimates
How it works.
Same pipeline, same enrichment, same multi-tenant data plane as the rest of the visualAI platform — extended with the studioGPT try-on engine and a bulk generation job for catalog-scale runs.
Eligibility
cleanerGPT flags every SKU as VTO-eligible based on garment category, image quality, and attribute completeness. Ineligible products are filtered upstream — shoppers never see a broken try-on, merchants never burn quota on a bad SKU.
Generate
The studioGPT engine combines the person image with the product image to produce a photoreal try-on. Bulk jobs run asynchronously per-merchant with per-tenant rate limits.
Distribute
Consumer outputs render inline on the PDP and optionally attach to the cart line. Bulk outputs land in your product media library — catalogGPT picks them up automatically on the next feed cycle and pushes them to every AI agent and search destination you syndicate to.
Why studioGPT, not a standalone VTO vendor.
The model itself is commoditizing fast. The platform around it isn't.
One engine, two surfaces
Most VTO vendors are either consumer-side widgets or merchant-side photo tools. studioGPT is one model, one ingestion path, one billing line — the same enrichment that lets shoppers try it on also generates the model shots that go in your catalog.
Purpose-built for apparel
Not a generic image generator with a wardrobe prompt. The studioGPT engine is tuned specifically for tops, bottoms, and one-pieces — with retry, idempotency, and observability matching the rest of the visualAI platform.
Multi-tenant from day one
Per-merchant row-level security on every query. Avatars, generated assets, and preview jobs are scoped to merchant_id at the DB layer — nothing crosses tenants, ever.
Diverse model library, opt-in custom
visualAI maintains a consented stock library covering body types, ethnicities, ages, and poses. Merchants can opt in to the library, upload their own model set, or let shoppers generate avatars from a single photo.
Common questions.
Which garments are supported?+
Apparel at launch — tops, bottoms, and one-pieces (dresses, jumpsuits). Shoes layer on independently. Accessories and outerwear with structure (coats, jackets) are on the roadmap.
What about privacy?+
Person images are session-scoped by default and not persisted unless a shopper explicitly opts in to "save my fit profile." Storage uses signed URLs with ≤ 5-minute expiry, region-pinned per merchant (EU merchants get an EU region), and every access is audit-logged. GDPR/CCPA delete cascades through both person images and generated outputs.
How does this fit with my existing photography?+
It replaces the on-model shoot line item, not the full pipeline. You still send flat lays or ghost mannequin shots — studioGPT turns those into on-model imagery across a diversity matrix you pick. The output replaces what you would have paid a studio to shoot.
What does it cost?+
studioGPT is a metered premium add-on inside shopperGPT (consumer try-on) and cleanerGPT (bulk catalog). Bulk is priced per-SKU at full margin; consumer try-on is per-event with a default sample_count of 2. Per-merchant monthly budgets and hard caps are configurable per plan tier.
Try it on. Or shoot the whole catalog.
The interactive MVP demo is available now. Book a call to see the catalog-scale generation pipeline and discuss early access.