GiftFit · One script tag · any store
The gift assistant for any store
Shoppers buying for someone chat with Mark-kun, the person they're buying for tries pieces on privately, and the giver buys the one that fits. The surprise stays a surprise.
See it run, right here
This one tag is the whole integration. Pick the options, then press Run: the same script loads on this page and Mark-kun appears in the bottom-right corner.
Language
Automatic: the store page's language, then each shopper's own (10 languages). No attribute needed.
Floating assistant data-widget
<script src="https://giftfit-469920173665.us-central1.run.app/giftfit.js" data-store="aurum" async></script>What the assistant does
Help me choose
The giver says who it's for, the occasion and a budget. The assistant shortlists pieces from your catalog and says why each one fits.
Private try-on
The recipient tries pieces on with an AI preview on their own phone. Their photos are deleted after the try-on and never shown to the giver or the store.
Fit Report
The giver gets fit scores, short reasons and the pieces that were hearted: enough to buy with confidence, without spoiling the surprise.
Voice, in the shopper's language
Shoppers can talk to the assistant and hear it answer. Voice stays off until they turn it on, and audio is never stored.
How it works
- 01
The giver asks: chat with the gift assistant on the store, set an occasion and a budget, and send a private link.
- 02
The recipient tries the pieces on privately with an AI preview on their phone and hearts their favourites.
- 03
The giver gets a Fit Report: scores, short reasons and what was hearted. The giver never sees their photos.
Quick start
Add GiftFit anywhere you can paste a script tag. On product pages the script reads the product from its JSON-LD or Open Graph tags; no catalog upload is needed to start.
Paste the tag before </body> on every page. Product pages show the product to the assistant; other pages get a general greeting. No language setting is needed: the assistant opens in the page's language and then follows each shopper's own.
<script src="https://giftfit-469920173665.us-central1.run.app/giftfit.js" data-store="aurum" async></script>| Attribute | What it does |
|---|---|
| data-store | Your store key. Default: aurum, the demo catalog. |
| data-lang | Optional. Fallback UI language, used when neither the page's lang nor the browser names an allowed one. Example: vi. |
| data-langs | Optional. Limits the assistant to these languages, comma-separated, e.g. en,vi. Without it, all ten are used: en,vi,ja,ko,zh,th,id,es,fr,de. |
| data-widget | "off" hides the floating assistant. |
| data-button | "off" hides the "Gift it right" button on product pages. |
In Shopify admin: Online Store → Themes → … → Edit code. Open layout/theme.liquid, paste the tag just before </body> and save. Shopify themes already set the page's language, so the assistant opens in the store's current language.
<script src="https://giftfit-469920173665.us-central1.run.app/giftfit.js" data-store="aurum" async></script>
</body>Easiest: install a header-and-footer code plugin and paste the tag into its footer box. Or print it from the wp_footer hook in a child theme's functions.php:
add_action( 'wp_footer', function () { ?>
<script src="https://giftfit-469920173665.us-central1.run.app/giftfit.js" data-store="aurum" async></script>
<?php } );WooCommerce product pages already include Product structured data, which the script reads.
In your container: Tags → New → Custom HTML. Paste the tag, set the trigger to All Pages, save, then Submit to publish.
<script src="https://giftfit-469920173665.us-central1.run.app/giftfit.js" data-store="aurum" async></script>Paste your store URL
GiftFit reads your public product data (structured data, sitemap, or the Shopify or WooCommerce product feed) and builds a private preview store with the assistant on it. Nothing is installed on your site, and the preview is deleted after 24 hours.
Only public product data is read and robots.txt is respected. The preview says clearly that it is not affiliated with your store.
For judges
No selfie needed: a seeded Fit Request with pre-recorded try-ons of an AI-generated model.