FOR AGENTS

UNDERCART MCP CONNECTOR

Query Undercart’s Shopify store intelligence from any MCP-compatible agent. Add one connector URL, sign in, and ask questions like “find beauty Shopify stores over $10M/yr that use Klaviyo” — no code to write.

Connect in 60 seconds
  1. 1
    In your MCP client, open its connector/server settings and choose to add a remote (HTTP) MCP server. In Claude that’s Settings → Connectors → Add custom connector; other clients call it “add server” or use an mcp.json entry.
  2. 2
    Name it Undercart and paste this connector URL:
    https://undercart.co/api/mcp
  3. 3
    Save. The client opens an Undercart sign-in window over OAuth — sign in with your Google account and click Allow access. The Undercart tools appear and you can start asking questions.

Requires a client that supports remote MCP servers over Streamable HTTP with OAuth. If yours doesn’t do OAuth, use the API-key header below instead. Note that adding a connector inside a Claude organization (rather than a personal account) typically needs a Team or Enterprise plan and org-admin access.

What you get

The connector exposes read-only tools over the Model Context Protocol (Streamable HTTP). Your agent can search stores, pull full profiles, list a store’s tech stack, find similar stores, view the actual ads a store is running, and read its captured marketing emails — across the whole Undercart catalog of Shopify stores. Every result reports estimated revenue annual before monthly and links back to the store’s Undercart page. It’s the same data as the REST API, driven by natural language instead of HTTP.

Alternative: API key

For MCP Inspector, scripted clients, or any tool that lets you set a custom request header, authenticate with an Undercart API key instead of OAuth. Create a key in Settings → API keys (shown once), then send it as an X-API-Key header on requests to https://undercart.co/api/mcp. Same key scheme as the Undercart REST API.

Tools

search_storesSearch Shopify storessearch

Filter the catalog with the Store Explorer’s filters: keywords in the store’s own description, category, country, annual revenue, traffic rank and trend, top market, catalog size and price, reviews, theme, hiring, email cadence, and apps installed, recently added, or recently removed. Returns concise profiles ranked by estimated revenue.

get_storeGet store profilestore lookup

Full profile for one store by domain: revenue and order estimates, Shopify plan, category, country, founding year, catalog stats, technology stack, ad presence, traffic rank and year-on-year trend, top markets, and social profiles.

list_store_technologiesList store technologiesstore lookup

The detected app/technology stack for a store, grouped by category (email, reviews, subscriptions, analytics, and more).

find_similar_storesFind similar storessearch

Peers in the same category and revenue band as a given store — useful for competitive sets and lookalikes.

get_store_adsGet store adsstore lookup

Per-platform ad presence (active, ad count, estimated spend) and the actual ads in a platform’s public ad library — headline, copy, CTA, landing page, and first/last seen dates for Meta, Google, and TikTok. In clients that support MCP Apps, results render as an ad gallery with image and video previews right in the conversation.

get_store_emailsGet store marketing emailsstore lookup

The marketing emails Undercart has captured from a store’s newsletter: subject, snippet, discount codes, and send cadence.

get_store_emailRead one marketing emailfree

The full body of a single captured email, by id — plain text plus the sanitized HTML source on request. In clients that support MCP Apps, the email renders visually, images included, right in the conversation.

list_store_categoriesList store categoriesfree

All store categories with store counts — the valid values for the search_stores category filter.

Example prompts

“Find beauty Shopify stores doing over $10M/yr that use Klaviyo.”
“Give me the full profile for allbirds.com — revenue, tech stack, and social following.”
“What apps and technologies does gymshark.com use?”
“Show me stores similar to allbirds.com in size and category.”
“List US apparel stores between $1M and $5M in annual revenue that run Recharge.”
“Which top-50k stores with rising traffic added Gorgias in the last 90 days?”
“Show me the actual ads gymshark.com is running on Meta right now.”
“Show me ridge.com's recent marketing emails — any discount codes?”

Usage & limits

Opening a store — get_store, list_store_technologies, get_store_ads or get_store_emails — is a store lookup, charged the first time that store is opened in a month. Opening it again that month is free, whichever tool asks and whether you opened it in ChatGPT, Claude, on Undercart or through the REST API. search_stores and find_similar_stores are searches, charged once per identical query per month. list_store_categories and get_store_email are free. On the free plan (25 store lookups and 50 searches a month), a lookup or a search through the connector or the REST API uses 5 of them, where the same thing on Undercart uses one; Unlimited ($29/mo) has no caps. When you reach a limit, tools return a message with the reset date and where the plans are described. Tools cover the whole catalog. Each store says whether Undercart has researched it in depth (enriched: true); one that hasn’t been researched carries public-web data only — apps, description, country and traffic rank — with no ads, emails or social yet, and the connector never starts that research itself. Pass enriched: true to search_stores for researched stores only. A store with no published revenue estimate shows a preliminary revenueBand from its popularity instead.

PREFER HTTP?
The same data over a plain REST API, with an X-API-Key header.
Read the API docs →
← Back to undercart.co · Questions? hello@undercart.co · Privacy · Terms · Support