Model Context Protocol
Connect Store Detector to Claude, ChatGPT, Cursor or n8n — then ask, in plain English, anything about 311,766 tracked Shopify stores. Traffic, revenue estimates, installed apps, pixels, themes and live detection. No dashboards, no exports.
Claude
ChatGPT
Cursor
Perplexity
Gemini
n8n
Demo
10 read-only tools over the same index that powers the sidebar and the public rankings. Your assistant calls them directly and answers with real figures.
Research
Ask how crowded a niche is, what it charges, which apps and themes it runs on, and who the biggest players are. The answer comes back from 311,766 tracked stores, not from the model's memory.
Prospect
Every store we track carries its installed apps and firing pixels. That makes “UK stores running Klaviyo with 50k+ visits” a single question — a qualified prospect list built on what a merchant already pays for.
Detect
detect_store fetches any URL live and reports whether it’s Shopify, which theme it runs, and which apps and pixels are on the page. The extension’s core trick, available to your agent.
Every tool is read-only. Your assistant picks the right ones for the question — you never name them yourself.
search_stores
Search the Shopify store index by country, category, installed app, pixel, traffic band and Shopify Plus status. Returns ranked matches with domain, category, traffic and revenue estimates.
Build a filtered list of stores that match a brief.
get_store
Full dossier for one store: platform, theme, apps, pixels, socials, category, price band, traffic and revenue estimates, and ad activity.
Everything the extension sidebar shows, as structured data.
top_stores
Leaderboard of the top Shopify stores by traffic, optionally filtered to a country and/or category.
Answer "who are the biggest players in X".
stores_using_app
Only here
List every store running a given Shopify app (Klaviyo, Recharge, Gorgias, Yotpo…), optionally scoped to a country.
Agency prospecting — a lead list filtered by tech stack.
stores_with_pixel
Only here
List stores running a given tracking pixel — Meta, TikTok, GA4 or Pinterest — optionally scoped to a country.
Find who is actually spending on a channel.
find_similar_stores
Find lookalike and competitor stores for a domain, same-market or international, from the precomputed similarity graph.
Turn one brand into a full competitive set.
compare_stores
Compare 2–5 stores side by side across traffic, revenue, apps, theme, price band and social following.
Competitive matrices without a spreadsheet.
market_stats
Aggregate market intelligence for a country and/or category: store counts, average price point, top apps and top themes.
Size a niche before you enter it.
store_growth
Social-follower growth history for a store across a look-back window, built from daily snapshots.
Separate the brands that are scaling from the ones that are flat.
detect_store
Only here
Live X-ray of any URL right now — is it Shopify, which theme, which apps and pixels. A real-time fetch, not a database lookup.
Works on stores that are not in the index yet.
One endpoint, every client. Chat apps sign in over OAuth; editors and automation tools take a config block or an API key.
MCP endpoint
https://mcp.storedetector.com/mcp
Streamable HTTP. Clients that still require the legacy transport can use
https://mcp.storedetector.com/sse.
Custom connector · OAuth
Developer mode · Plus, Pro, Team or Enterprise
One command in your terminal
claude mcp add --transport http store-detector https://mcp.storedetector.com/mcpmcp.json · also works in Windsurf & VS Code
{
"mcpServers": {
"store-detector": {
"url": "https://mcp.storedetector.com/mcp"
}
}
}API key · for automations and scripts
curl -N https://mcp.storedetector.com/mcp \
-H "Authorization: Bearer $STORE_DETECTOR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'Copy one, paste it into your assistant, then follow up in your own words. For the heavier briefs, use a reasoning model rather than a speed-optimised one.
A plain-English read on how crowded a market is, what it charges and what it runs on. Good first prompt to see the tools fire.
“Give me a market snapshot of the skincare niche in the United States: how many Shopify stores you track, the average price point, the most common apps and themes, and the five biggest stores by traffic. Flag anything that surprises you.”
The extension sidebar, as text you can paste into a doc — and it works on stores that are not in the index yet.
“X-ray gymshark.com for me. What theme and apps does it run, which pixels are firing, what is the estimated traffic and revenue, and what is the price band? Then tell me what that stack says about how they operate.”
A ranked list you can act on, without opening a dashboard or exporting a CSV.
“List the top 20 Shopify stores in Australia by traffic. For each one give me the category, estimated monthly visits and whether it is on Shopify Plus. Group them by category at the end.”
The competitor deck slide, generated from one domain — no manual list-building first.
“Find the closest competitors to allbirds.com, pick the six most relevant, and build me a competitive matrix comparing traffic, estimated revenue, price band, theme, apps and social following. Finish with three things Allbirds does differently from the pack.”
A qualified lead list filtered by what a store already pays for — the fastest signal that they buy software.
“Find UK Shopify stores running Klaviyo with more than 50,000 monthly visits. Give me domain, category, estimated traffic and what other apps they run. Sort by traffic and tell me which five are the best fit for a retention-marketing pitch, and why.”
Pixel presence is a spend signal. This separates the brands running paid social from the ones that only talk about it.
“Which stores in the home-and-garden category have a TikTok pixel installed? Show me the 25 largest by traffic, note which of them also run a Meta pixel, and summarise what that says about where the category is putting its budget.”
The research half of a launch decision — market size, incumbents, the standard stack and the gap — in one pass.
“I am considering launching a pet supplements brand in Canada. Build me an entry brief: how big is the tracked market, who are the top ten incumbents and how big are they, what apps and themes are standard, what price points cluster where, and which incumbents are growing fastest. End with the three most defensible gaps you can see and what evidence supports each.”
The stack the winners converge on, and the one thing only the top few do. Turns a list into a playbook.
“Take the top 20 menswear Shopify stores by traffic. Pull the full stack for each, then find the patterns: which apps appear in most of them, which themes dominate, which pixels are near-universal, and what only the top five do that the rest do not. Give me a stack recommendation for a new entrant and say where you are inferring rather than measuring.”
A Monday-morning briefing on a watchlist you define. Replaces the manual round of tab-opening.
“Here is my watchlist: [domain1], [domain2], [domain3]. For each, pull the current dossier and the social-follower trend over the last 90 days. Tell me who is growing, who has stalled, and any stack changes worth noticing. Format it as a short briefing I can paste into Slack, with the numbers inline.”
Connect the Store Detector MCP server to Claude, ChatGPT, Cursor or n8n.
Copy any prompt above and paste it into the conversation.
Follow up naturally — narrow the filters, ask for a different cut, or ask it to show its working.
Worth knowing before you connect it, not after. These are real gaps in our dataset, and where a competitor covers one we say so.
We report whether a store is running Meta ads and how many, not the creatives themselves. If you need a swipe file of ad copy and video, TrendTrack or BrandSearch cover that and we do not.
We detect the email platform a store runs — Klaviyo, Omnisend, Attentive — but we do not archive the campaigns they send.
The index is Shopify. WooCommerce, BigCommerce, Amazon and TikTok Shop are out of scope, though detect_store will still tell you what an arbitrary URL is built on.
Traffic comes from third-party panel data; revenue is modelled from traffic and category conversion benchmarks. Directionally useful, never accounting data — the methodology page shows the workings.
Full detail on how the numbers are produced is on the methodology page.
Agent access is included with Store Detector Pro under a fair-use ceiling — one subscription covers the extension and the MCP server. The two competing store-data MCP servers charge $99–$349 a month for agent access on top of their app subscription.
See pricingPart of the data is already public, free and unauthenticated — no MCP client required.
The Model Context Protocol is an open standard for connecting AI assistants to external data and tools. Once the Store Detector server is connected, your assistant can query our Shopify store index directly and answer with real figures instead of guessing — no copy-pasting, no exports.
Anything that speaks MCP: Claude (web, desktop and Claude Code), ChatGPT in developer mode, Cursor, Windsurf, VS Code, Perplexity, Gemini CLI and n8n. Claude, ChatGPT and Perplexity connect over OAuth — paste the URL and sign in. Editors and automation tools take a JSON config or an API key.
Yes — one index behind both. The MCP server reads the same 311,766 tracked Shopify stores that power the extension sidebar and the public rankings, plus live detection for URLs that aren't tracked yet.
No. Every tool in the current surface is read-only. Your assistant can query, compare and summarise, and produce artifacts you use elsewhere — it cannot change anything on our side.
Queries are answered in real time and we do not store conversation content. We log which tool was called and when, for quota and abuse protection. Anything your assistant keeps is governed by that assistant’s own retention policy, not ours.
MCP access is bundled into Store Detector Pro rather than sold separately — one subscription covers the extension and agent access under a fair-use ceiling. Competing store-data MCP servers price agent access from $99 to $349 a month on top of their app subscription.
No, and we would rather say so here than after you subscribe. We report whether a store runs Meta ads and which email platform it uses, but we do not archive ad creative or email campaigns. See “What it does not do” above for the full list of gaps.
Yes, and part of it is already public and free: /openapi.json describes read-only rankings endpoints that need no authentication. Every page also has a markdown twin on Accept: text/markdown, and /llms-full.txt carries the full ranking data.