Shopfront Concierge
Shopfront Concierge is a shopping assistant for a made-up homeware shop called Kettle & Loom. It does not guess: it calls tools to search the catalogue, look up an order and recommend products, and its answers quote product details back from the catalogue. The shop has an admin view and an evaluation page that scores how well the assistant picks the right tool.
Measured numbers
- Tests
- 507
- Browser tests
- 59
- Lighthouse performance
- 93
- Accessibility
- 100
- Best practices
- 100
- SEO
- 100
Home page, mobile. Copied from the showcase's own README on 24 October 2026.
Open the live site (opens in a new tab) shop.phoenixcodes.org
On a phone
At desktop width
What it proves
- AI tool calling with checked arguments, so answers come from live data and not from memory.
- A working shop around it, with catalogue, cart and checkout, so the assistant is tested in context.
- Honest limits: it holds back when the catalogue has no match.
- An admin view and an evaluation harness measure search quality and tool choice.
How this adapts to your project
Connect the tools to your catalogue and order system, whether that is Shopify, WooCommerce or your own API, and the assistant works on real stock and real orders. The evaluation set is where you write down the questions your customers actually ask.