You can connect Stripe to Claude in a few minutes using the Model Context Protocol (MCP), and once it’s connected, Claude can look up a customer, check a payment, or walk through a subscription without you digging through the Stripe dashboard by hand. This guide covers both ways to do it: Stripe’s official remote MCP server connected directly, and connecting it through a gateway if you use more than one AI client.

What the Stripe MCP server actually does

Stripe publishes an official remote MCP server at mcp.stripe.com. It exposes your account’s customers, payments, and subscriptions as tools Claude can call directly, so prompts like these work against your live Stripe data:

  • “Look up the last three payments from this customer.”
  • “Is this subscription still active, and when does it renew?”
  • “Find the customer tied to this email and summarize their billing history.”
  • “Issue a refund for this charge.”

Claude acts through the same permissions your Stripe account already has — it isn’t a separate export or a script you maintain, it’s your account looking things up, and in some cases changing them, on your behalf.

Option 1: Connect Stripe to Claude directly

Claude supports remote MCP servers natively. The short version of the setup:

  • Open Claude’s connector or MCP settings.
  • Add a remote server pointing at Stripe’s MCP endpoint.
  • Sign in with your Stripe account when prompted — this is standard OAuth, so Claude never sees your Stripe password or your secret API key, and you can revoke access from Stripe’s own dashboard at any time.

That’s enough if Claude is the only AI client you use with Stripe. The tradeoffs show up once you add a second tool or a second teammate: every additional AI client (ChatGPT, Cursor, an agent) needs its own copy of that same connection, its own OAuth grant, and its own place where the config can go stale.

Option 2: Connect it once, through a gateway

If you’re already juggling more than one MCP server or more than one AI client, wiring each pair directly gets old fast — that’s the connection sprawl we wrote about in what an MCP gateway is. The alternative is connecting Stripe (and everything else) to one gateway URL, then pointing every AI client at that single URL instead of reconfiguring each one.

With gate specifically, connecting Stripe looks like this: add gate’s gateway URL to Claude once, then pick Stripe from the Stripe server page in the directory and sign in with OAuth the same way you would directly. From then on, Stripe’s tools are available in every client connected to your gate URL — Claude, ChatGPT, or Cursor — without a second setup step. See the full list of servers set up this way on the MCP servers page.

Why the connection method matters for a payments account

A Stripe account is one of the more consequential things you can hand an AI client access to — it holds customer records, payment history, and the ability to move money back out through a refund. A few things worth checking before or right after you connect:

  • Is it the official endpoint? Connect Stripe’s documented server (mcp.stripe.com), not a third-party mirror claiming to proxy it.
  • Read, or read and write? A connection that can look up a customer can also issue a refund — that is not the same risk, and it is worth knowing which tools are read-only before you hand an assistant a prompt like “clean up duplicate charges.”
  • Test mode first. Stripe accounts have a live and a test mode with separate data; trying prompts against test-mode data before pointing Claude at live customer records is a reasonable way to see what a tool call actually returns.
  • Who else on the team can reach it? If more than one person connects Stripe to their own AI client, a shared, revocable setup beats everyone holding their own grant — the same access-control question covered in MCP access control for teams.

You can run a quick, no-signup check on any server’s tool list with gate’s free MCP security scanner, which looks for prompt injection and hidden instructions in tool descriptions before you rely on a server day to day.

Troubleshooting the connection

A few things that commonly go wrong when connecting any remote MCP server, Stripe included:

  • OAuth loop or sign-in fails. Make sure you’re signed into the right Stripe account — and the right account, if you manage more than one business — before authorizing.
  • Claude can’t find a charge you know exists. Stripe’s tools respect your account’s permissions and mode; a charge made in live mode won’t show up while Claude is connected to test mode, or the other way around.
  • Tools disappear after a Stripe update. Remote servers can change their tool list without warning. If you’re connected through a gateway, that kind of change is exactly what drift detection is meant to catch — see how MCP rug pulls work for why a one-time review isn’t enough.
Where gate fits: gate connects Stripe once, behind a URL every AI client can share, with allow/ask/block rules you can set per tool — so a lookup can run automatically while a refund waits for a human. Every server is scanned for prompt injection before it goes live, and every call is recorded in a plain-language log. See the full gateway at /mcp-gateway.

The bottom line

Connecting Stripe to Claude with MCP takes a few minutes either way: point Claude directly at Stripe’s server and sign in, or connect it once through a gateway if you want the same access available in every AI client you use, with rules on what it can do without asking first. Either way, a payments account is exactly the kind of connection worth a minute’s thought before you say yes — especially the part that can move money.