Retrieves live product and collection data from a Strapi v4-powered headless CMS web-builder plugin, using Strapi filters, sorting, pagination, and Bearer token authentication to search products and fetch product details.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Crunchy Fashion MCP server show me the newest earrings under ₹1500 in the jewellery collection"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server (built with mcp-use) that searches live products on crunchyfashion.in and renders them as a shoppable carousel styled like the storefront. Clicking a product opens its page on the website.
Products come from the headless CMS web-builder plugin:
GET {CRUNCHY_API_URL}/api/web-builder/products/{CRUNCHY_WEBSITE_NAME} (Strapi v4 filters,
sort and pagination, Bearer token auth).
npm install
cp .env.example .env # then fill in CRUNCHY_API_URL and CRUNCHY_API_TOKEN
npm run devOpen http://localhost:3000/mcp/inspector to try the tools.
Variable | Required | Default |
| yes | – (the CMS gateway, e.g. |
| yes | – (Strapi API token with read access to products and collections) |
| no |
|
| no |
|
Related MCP server: Agentic Product Protocol MCP Server
search-products (renders the product-search view): keyword query, collection,
minPrice / maxPrice (INR), sort (newest, price-asc, price-desc, name-asc),
page, pageSize (max 24). Only live products are returned (isLive true or unset).
Keywords must all match the product title/SKU first; if nothing matches, the search widens
to descriptions, then to products matching any keyword. Common spelling variants
(jewellery/jewelry, earrings/earings, mehndi/mehendi…) are handled.
get-product-details: description, attributes, shades, FAQs and price for one product
by slug or SKU.
list-collections: live collection names and handles to use as the collection filter.
views/product-search/ — Montserrat on white, black CTAs and the magenta % OFF pill from
the storefront. Inline it is a swipeable carousel with a "Show more" tile; in fullscreen it
becomes a product grid with "Load more". Shade swatches switch the card image and price.
Images are served through the storefront's /_next/image optimizer (small AVIF/WebP) and fall
back to the original CDN file. Product links carry utm_source=mcp&utm_medium=ai_assistant.
npm run typecheck # type-check server + views
npm run build # production bundle in .mcp-use/build
npm run start # serve the production build
npm run deploy # deploy to Manufact CloudThis server cannot be deployed
Search and get fashion products recommendations across multiple e-ecom stores
Search products, view details, and try items on across stores powered by Tolstoy.
AI shopping gateway for product search, inventory, carts, and merchant-hosted checkout.
Visual fashion search across retailers and resale, with localized prices and app handoffs.
