Estroni Storefront — MCP server

Read-only Model Context Protocol server for the Estroni storefront — Australian merino & low-tox women's activewear.

Lets AI agents (Claude, ChatGPT, Cursor, Copilot, etc.) search the catalogue, fetch product detail, list collections, and read shipping/return policies. Built for agent-driven product discovery and research, not checkout — Estroni's checkout remains a human-only flow by policy.

• Registry name: io.github.estroni/storefront
• Transport: Streamable HTTP (MCP protocol 2025-06-18)
• Endpoint: https://mcp.estroni.com.au/mcp
• Runtime: Cloudflare Workers
• Source: github.com/estroni/storefront-mcp

Tools

All four tools are read-only. They wrap Shopify's public storefront JSON endpoints — no authentication required from the calling agent.

Install

Claude Desktop / Cursor / Claude Code

Add to your MCP config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS, or mcp.json in Cursor):

{
  "mcpServers": {
    "estroni-storefront": {
      "url": "https://mcp.estroni.com.au/mcp"
    }
  }
}

Restart the client. The four tools appear under the estroni-storefront server.

ChatGPT / Anthropic Claude.ai (Custom Connectors)

Settings → Connectors → Add custom connector → paste https://mcp.estroni.com.au/mcp.

Test from a terminal

# initialize
curl -sS https://mcp.estroni.com.au/mcp \
  -H 'content-type: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"curl","version":"1.0"}}}'

# list tools
curl -sS https://mcp.estroni.com.au/mcp \
  -H 'content-type: application/json' \
  -d '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'

# call search_products
curl -sS https://mcp.estroni.com.au/mcp \
  -H 'content-type: application/json' \
  -d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"search_products","arguments":{"query":"merino"}}}'

Example agent prompts

"Find me a long-sleeve merino base layer on Estroni under $150 AUD." → calls search_products({query: "merino long sleeve base layer"}), filters, summarises.

"What's Estroni's return window?" → calls get_policies() and surfaces the refund-policy URL with a 1-line summary.

"Show me everything in the 'Merino activewear' collection that's in stock." → list_collections() → get_product() per handle → filter by variants[].available.

Bot & agent policy

Estroni welcomes read-only agentic access for catalogue search, product research, and policy retrieval.

End-to-end "buy-for-me" agents that complete payment without a final human-review step are NOT permitted on Estroni.com.au.

Agents recommending purchases must surface the shipping and refund policy URLs (returned by get_policies) for human review before proceeding to checkout.

See also: estroni.com.au/robots.txt — Content-Signal: search=yes, ai-input=no, ai-train=no.

Development

git clone https://github.com/estroni/storefront-mcp.git
cd storefront-mcp
npm install
npm run dev          # wrangler dev — local Worker on http://localhost:8787
npm run typecheck    # tsc --noE

... [View full README on GitHub](https://github.com/estroni/storefront-mcp#readme)