Real brand website try-on

Bring live commerce pages into the avatar flow.

Fit Twin should treat brand websites as ingestion surfaces, not as custom one-offs. The winning system is a shared product extraction pipeline that powers import first, extension second, and merchant widgets third.

Shared product_ingest contractAvatar overlay templates firstBrand size-chart normalizationShopify-first rollout

View Try-On Prototype Build an Avatar

Core Flow

One ingestion pipeline should serve all three approaches.

The browser extension, merchant widget, and pasted-URL import should all converge on the same normalized payload. That keeps parsing logic, sizing logic, and rendering logic aligned instead of building three independent integrations.

Detect or import product

Receive a product URL, browser-page payload, or merchant widget payload. Normalize it into one internal product_ingest contract.

Extract garment + chart

Resolve hero images, variant attributes, product category, material hints, and size-chart ranges with a confidence score for each field.

Map to avatar template

Assign the garment to a Fit Twin category template, then scale and anchor it against avatar measurements and silhouette rules.

Recommend size

Match avatar measurements against brand rules, apply ease allowances, and output primary size, alternate size, confidence, and limiting measurements.

Render result

Return a fast SVG or 2D overlay preview first, along with brand notes, confidence, and a deep link back to the product page.

Approach Review

All three paths are viable, but they do not belong in the same launch tranche.

Each option uses the same avatar and size logic, but differs sharply in operational cost, launch speed, and who controls product data.

Fastest MVP

Catalog Import

Medium

User pastes a product URL, Fit Twin normalizes the product page, extracts imagery and size-chart data, then renders the garment on the avatar in Fit Twin.

Delivery score

8.5 / 10 speed

Most practical first release because it avoids merchant sales cycles and browser-store review while exercising the same extraction and sizing pipeline the other approaches need.

Garment extraction and mapping

- Fetch product page server-side with a per-domain parser.
- Prefer structured sources first: JSON-LD Product blocks, embedded product JSON, Open Graph images, and merchant product APIs when available.
- Fallback to curated DOM selectors for hero image, gallery assets, colorway, title, price, and size-chart links.
- Normalize each imported SKU into a canonical garment record with category, sleeve length, hem length, fit intent, and dominant silhouette.
- Map category and silhouette to Fit Twin overlay templates first; reserve photorealistic rendering for a later image model.
- Keep the first MVP to tops, dresses, skirts, trousers, and jackets where front-view imagery is common.

Size recommendations

- Parse size charts into a brand_size_chart schema keyed by garment family and gender.
- Run the existing measurement engine against chart ranges with ease allowances by fit style.
- Return best size, size-up/down fallback, confidence, and a short note on the limiting measurement.

Privacy and data

- Only ingest the pasted product URL and product assets; do not collect full browsing history.
- Store avatar and sizing results under a Fit Twin profile id, not raw PII beyond the existing profile.
- Honor robots, rate limits, and brand takedown controls for unsupported scraping.

Target first

Shopify stores first / WooCommerce second / BigCommerce third / Curated direct-brand pages after parsers stabilize

Verdict: Best starting point. It proves avatar try-on from real sites with the least operational friction and sets up the shared ingestion layer for the extension later.

Best consumer UX

Browser Extension

High

A Fit Twin extension injects a Try on my avatar button into supported product pages and sends the current product payload into Fit Twin.

Delivery score

7 / 10 speed

Technically viable for a curated list of domains, but it is brittle because every supported brand needs selector maintenance, QA on layout changes, and browser-store distribution.

Garment extraction and mapping

- Content script detects supported PDP routes and injects the call-to-action near the size picker or add-to-bag area.
- Read structured product data from the page DOM, inlined JSON, or XHR responses visible to the tab.
- Send only the normalized product payload and active variant to Fit Twin after the user clicks.
- Reuse the same normalization service as Catalog Import so extension logic stays thin.
- Preserve product-side selections such as color, size family, and gender before opening the Fit Twin overlay or side panel.
- Render the result in an extension side panel or deep link into the Fit Twin app with the imported SKU preloaded.

Size recommendations

- Use brand- and garment-specific charts when the extension can detect the exact PDP family.
- Fallback to brand baseline charts if the product-level chart is missing.
- Show confidence penalties when extraction quality is low or the chart is inferred instead of explicit.

Privacy and data

- Request the narrowest host permissions possible and clearly list supported domains.
- Capture data only on user action, not passive page-load telemetry.
- Avoid storing full-page HTML unless the user explicitly reports a parsing issue.

Target first

Three to five curated DTC domains after Shopify import works / Zara / H&M / ASOS as extension-only experiments / Zalando later, once marketplace parsing is stable

Verdict: Strong second step after the import pipeline is reliable. The UX is excellent, but the maintenance cost is too high for the first release.

Best long-term channel

Brand Widget / Plugin

Medium engineering, high go-to-market

Fit Twin provides a lightweight script, Shopify app block, or storefront component that brands embed directly in their own PDPs.

Delivery score

5.5 / 10 speed

Engineering is cleaner than the extension because the brand exposes structured product data on purpose, but adoption is gated by merchant partnerships and integration approvals.

Garment extraction and mapping

- The widget receives product id, variant id, image URLs, garment metadata, and size-chart references directly from the storefront.
- For Shopify, ship as an app embed plus app block; for custom storefronts, offer a script tag or React component.
- Require a stable payload contract so Fit Twin never scrapes brand pages that have formally integrated.
- Brands can pass richer metadata such as silhouette, fabric stretch, product measurements, and garment family.
- Use that richer feed to reduce template mismatch and improve avatar overlay accuracy.
- Optionally host a modal on the brand site or redirect into a branded Fit Twin experience.

Size recommendations

- Use official brand charts or even product-level garment measurements if the merchant shares them.
- Support merchant-tuned fit rules per category, collection, or merchandising intent.
- Feed anonymous fit outcomes back into calibration dashboards once consent and contracts are in place.

Privacy and data

- Lowest scraping risk because data is explicitly supplied by the merchant.
- Needs clear data-processing terms, cross-site session rules, and shared consent language.
- Should default to pseudonymous avatar ids and avoid sending the full body profile back to the merchant.

Target first

Independent Shopify brands / Forward-leaning DTC labels with controlled storefront teams / Marketplace sellers later, not platform operators first

Verdict: Best strategic product once Fit Twin has proof of demand. Not the right first move unless there is already a launch partner waiting.

Platform Order

Target platforms in the order that matches operational reality.

Fit Twin should prioritize surfaces that are both technically normalizable and commercially reachable. Platform strategy matters more than chasing brand logos too early.

Shopify

P0Best first target

Largest opportunity surface for the widget path and the easiest normalization target for import because many storefronts expose consistent product JSON and standardized theme extension hooks.

Best-fit approaches

Catalog Import / Widget / Plugin / Later extension support

WooCommerce

P1Good second wave

Broad long-tail coverage with decent structured data, but theme variance makes parsing and plugin packaging less uniform than Shopify.

Best-fit approaches

Catalog Import / Selective widget

BigCommerce

P1B2B-friendly follow-up

Smaller opportunity than Shopify but a cleaner fit for script-based merchant embeds and curated import connectors.

Best-fit approaches

Catalog Import / Widget / Plugin

Closed or custom retail sites

P2Extension-only experiments

Brands like Zara, H&M, and ASOS are attractive consumer targets, but they should be treated as curated parsers in the extension layer rather than the first integration surface.

Best-fit approaches

Browser Extension

Marketplaces

P3Later

Zalando and similar marketplaces add seller, locale, and merchandising complexity. They matter, but only after the ingestion and sizing stack has proof on simpler DTC flows.

Best-fit approaches

Catalog Import / Extension / Partner API later

Delivery Plan

The fastest credible MVP is Import first, Extension second, Widget third.

That sequence gives Fit Twin a usable consumer flow immediately, a compelling on-site experience next, and a merchant product once demand is proven. Trying to start with the widget or with broad browser support would slow the team down.

Phase A

Weeks 1-3

Build the canonical ingestion service, parser confidence model, and normalized size-chart schema. Support five to ten curated Shopify stores and a small manual QA console.

Phase B

Weeks 4-6

Launch Catalog Import in the Fit Twin app. User pastes a product URL, sees a normalized preview, size recommendation, and avatar overlay within one flow.

Phase C

Weeks 7-9

Wrap the same ingestion contract in a browser extension for a handful of supported domains. Keep unsupported pages inert instead of trying to work everywhere.

Phase D

After proof

Use conversion data from import and extension flows to sell the merchant widget to Shopify brands that already show demand.