Two real Australian stores, running on ShopFlare right now.
One sells vintage motorcycle parts with a deeply structured fitment catalogue. The other sells quilting and craft supplies where the search has to read product photos. Different shapes, same engine. Every query below runs against a live production catalogue.
Vintage motorcycle parts. Hard-shaped catalogue.
This is the search problem the incumbents can't solve cleanly. Customers know the year and model of their bike, not the part number. They know an old British name for a part, not the canonical SKU title. They know the photo of the part from the workshop manual, not the brand we stock it under.
ShopFlare's fitments index links every part to the bikes it fits, the trigram index matches any partial OEM number, and alias-aware search resolves "contact points set" to "Lucas Contact Set, Twins 1968-72". The MCP server lets AI assistants answer fitment questions directly.
- Products indexed
- 5,500+
- Motorcycle fitments
- 343
- Models covered
- 13
- Autocomplete time
- ~2ms
- MCP tools exposed
- 13
- Cloudflare cost
- ~$5/mo
Quilting & craft supplies. Search that reads the photo.
Birch Creative sells threads, battings, rotary cutters and haberdashery — thousands of products where the thing the customer is looking for is often a colour, a fibre, or a pack size that never made it into the product title.
ShopFlare's vision pipeline runs Workers AI over every product photo and appends what it sees — "white", "100% wool", "queen size" — to the searchable body. So "white quilting thread" or "100% wool batting" finds the right product even when the title only says a SKU. They moved off Doofinder and saved around $200/month.
A generic search box only knows the words in the title. ShopFlare reads the picture too — colour, material, pack count — and makes all of it searchable. That's the difference between "no results" and a sale.
Want yours next?
Send us your WooCommerce URL and we'll have a sandbox running inside 48 hours.
Request a sandbox