Your App, Now on Voice Command: Introducing GoodBarber's MCP Server
Written by Pierre-Laurent Medori on
Manage your entire app — products, orders, promos, push notifications — by talking to your AI assistant. One sentence, zero friction.
The two-minute jobs that don't need a laptop

If you run a GoodBarber app, you know the feeling. You're between two meetings and a supplier texts you that a product is back in stock. Or you're on the train and a flash sale idea pops into your head. Or a customer DM reminds you that yesterday's batch still hasn't shipped.
None of these are complicated tasks. They're two-minute jobs — the kind of thing that should take less time to do than to think about. But in practice, they require opening your laptop, logging in, navigating to the right section, and making the change. By the time you've done all that, the moment has passed, or you've added it to a mental to-do list that keeps growing.
That's the gap we wanted to close. Not by redesigning a screen or adding a shortcut, but by letting you act on your app the same way you'd ask a colleague for help — by describing what you need in a sentence. We're launching GoodBarber's MCP server, and it changes the way you interact with your app.
What is MCP, and why should you care?
MCP stands for Model Context Protocol. If that sounds technical, don't worry — the concept is straightforward.
Think of it this way: AI assistants like Claude, ChatGPT, or Cursor are incredibly good at understanding what you mean when you type a request in natural language. But on their own, they can't do anything on your app. They don't have access to your products, your orders, or your customer list. They're smart, but disconnected.
MCP is the bridge. It's an open standard — not proprietary to GoodBarber or any single company — that allows an AI assistant to securely connect to an external service and perform real actions on it. When you plug GoodBarber's MCP server into your AI assistant, you're essentially giving it a direct line to your app's data and operations. The assistant can now read your catalog, create a promo code, send a push notification, or pull your analytics — all because MCP provides the structured connection between what you say and what your app does.
The important word here is open. MCP isn't a GoodBarber invention. It's a protocol supported by a growing ecosystem of AI tools. That means you're not locked into one assistant — if you use Claude today and switch to something else tomorrow, your MCP server still works. You bring your own AI, and GoodBarber provides the connection to your app.
From Zapier MCP to GoodBarber's own MCP server
If you've been following our blog, this might sound familiar. Back in July 2025, we published an article exploring how to pilot a GoodBarber app through Zapier's MCP server. The idea was the same — manage your app by talking to an AI assistant — and it worked. We demonstrated sending push notifications and batch-updating tracking URLs through Claude, using Zapier as the relay between the assistant and your app.
But that setup had a structural limitation: the assistant could only do what Zapier exposed. If an action or a trigger wasn't declared on Zapier's side, the assistant simply didn't know it existed. You were working with a subset of your app's capabilities, filtered through a third-party catalog.
With GoodBarber's own MCP server, the connection is direct. No intermediary, no subset. The assistant connects straight to your app and accesses the full scope of operations — products, variants, collections, orders, promo codes, push notifications, customer data, loyalty, prospects, analytics. Everything that's available in your dashboard is now available in conversation.
The Zapier approach was a great proof of concept and remains useful if you want to chain GoodBarber actions with other services (a new order triggers a Slack message, a form submission creates a customer — that kind of cross-platform automation). But for managing your GoodBarber app directly, the native MCP server is faster, more complete, and requires no third-party configuration.
What you can actually do
This isn't a chatbot that answers questions about your app. It operates your app. Here's the full scope of what's available today:
- Shop management — create, update, and organize products, variants, collections, slides, PDFs, and media. Need to add a new product with three size variants and upload its images? One conversation.
- Order operations — track orders, update shipping status, check fulfillment. Instead of scrolling through a list, ask your assistant to surface exactly what you need — pending orders, today's shipments, a specific customer's history.
- Marketing — launch promo codes based on order amount, specific collections, individual products, or tags. Send broadcast push notifications to all your users or targeted segments. The kind of coordinated action that usually requires hopping between two or three sections, done in a single request.
- Customer intelligence — search your customer base, manage loyalty points, review prospects. Want to know who your top 5 buyers were last month? Ask.
- Analytics — pull page views, downloads, session times, and unique launches, broken down by platform, time period, or day of the week. Your numbers, delivered in conversation, ready to act on.
Every action is scoped to your app. The assistant doesn't have free rein — it accesses only what you've authorized, and every sensitive operation (sending a push to all users, modifying pricing, shipping orders) requires your explicit confirmation before it runs. You stay in control. You just stop navigating.
A new rhythm for managing your app
The real shift isn't about speed, though it is faster. It's about when and where you can manage your app.
Right now, most app management happens in dedicated sessions — you sit down, open the dashboard, batch your updates, and move on. That works well for planned work. But a lot of app management isn't planned. It's reactive. A restock arrives, a customer complains, a marketing opportunity appears. These moments don't wait for you to be at your desk.
With MCP, your assistant becomes the natural extension of your app. You're already in a conversation with Claude or Cursor for something else — drafting an email, brainstorming a campaign, reviewing your numbers — and you just add a line. "While you're at it, create a 15% code for that new collection and push it to my VIP customers." The boundary between thinking about your app and acting on your app disappears.
And because the AI checks context before acting and confirms results after, you always know what happened. No more wondering if the promo went live or if the push actually sent. The assistant reports back, every time.
Getting started
Setting up takes about three minutes:
- Add the MCP server URL to your AI client. This works with Claude, Cursor, and any other tool that supports the MCP protocol.
- Sign in with your GoodBarber credentials.
- Ask your first question:"What can you do on my app?" — the assistant will list every available action, so you know exactly what's possible.
No local install. No scripts to maintain. No developer account needed. If you can type a sentence, you can manage your app.
What's next
MCP is a starting point — but it's a starting point with compounding returns.
As AI assistants improve — better reasoning, wider context windows, more reliable execution — your app gains capabilities automatically. You don't update anything. You don't install anything. Every improvement to the AI models translates directly into a more capable assistant for your app. The same MCP connection that lets you create a promo code today might let you run a full marketing campaign analysis tomorrow, without any change on your side.
We've been building tools for app creators since 2011. The constant has always been the same: give people more power with less complexity. MCP is the next step in that direction.
Your app just learned to listen.
Connect your app today and start talking to your shop.
Design