Ask your AI assistant what's in stock, what to send and what it will cost. It prepares the quote. You approve it. Nothing ships until a person says yes.
Companion guide: agentic commerce for branded merch.
Sunday MCP is the Model Context Protocol server for Sunday's merchandise platform. It connects your Sunday account to Claude, ChatGPT, Gemini or any other MCP client, so you can check stock, build budgets, request designs, plan shipments and reorder merch in plain language. The assistant prepares quotes. A person approves every order and shipment.
The short version: whatever you do in the Sunday platform today, you can now ask for in the AI assistant you already use. The creative work gets faster. The commitments stay human.
What is Sunday MCP?
Sunday MCP connects Sunday's whole merchandise platform to the AI assistant on your screen. It's live. It works with Claude, ChatGPT and Gemini, and with any other assistant that speaks MCP. Some people search for it as Sunday for Claude or Sunday for ChatGPT. Same thing, one connector.
MCP stands for Model Context Protocol. It's an open standard that lets an AI assistant use another piece of software for you: read its data, run its actions and bring the result back into the chat. If the term is new, our plain-English guide to MCP servers covers it in a few minutes. And if your IT team asks why this is different from the integrations they already run, point them to MCP vs API.
In practice it means your catalog, your stock, your quotes, your shipments and your reporting are one sentence away. You don't open another tab. You don't scroll through a full catalog. You describe what you need and who it's for, and the assistant does the legwork inside Sunday.
Here is why we built it. The Sunday platform already takes care of the slow, physical part of merch: production, warehousing and shipping to 200+ countries. What still eats an afternoon is the thinking around it. Which products fit this audience. How to split a budget. What to reorder before the next event. That's work an AI assistant is good at, as long as it can see your merch.
What can you ask Sunday MCP?
Anything you'd do in the platform. Niels Vandecasteele at Sunday puts it this way: "How many hoodies do we still have in stock? Create me a quote to reorder the pants." Two sentences, two jobs that used to mean logging in, finding the right product and filling in a form.
Illustrative example · You, in Claude: How many black hoodies do we have left in stock, by size? If any size is running low, draft a reorder quote.
Sunday MCP: Here's your current stock for the black hoodie, split by size. Two sizes are well below the rest. I've drafted a reorder quote for those sizes, based on your last order of this hoodie. Open the quote link to review and approve it. Nothing is ordered until you do.
That covers the most common request. The rest of what you can ask falls into a handful of jobs.
Stock and data questions
Current stock by product and size. How many items shipped this quarter. Which countries you ship to most. These used to mean an export, a spreadsheet and a pivot table. Now they're a question, answered from your own inventory in Sunday.
Stock questions get specific fast: which colour, which size, how many left. Asking the assistant before an item runs out is the easiest habit to build.
Reorders, as quotes
One prompt, one quote. The assistant looks at what you ordered before, builds the quote and shows it to you. You approve it through a link. For a repeat item with an approved design, that is the whole process.
Budgets and wish lists
Ask Niels which task he wanted to kill first and he'll say assembling budgets. "You have 20k to spend and a target audience." With one prompt, Sunday MCP searches the whole catalog, looks at what's popular and what performs best, and comes back with a budget split. It can turn that into a wish list you share with colleagues before anyone commits to anything.
You, in ChatGPT: We have 20k for merch for our sales kickoff in Lisbon. Around 250 people, mostly remote sellers. Suggest a budget split and make a wish list I can share with our events lead.
Product and kit suggestions
This is where the real value sits. Time saving was never the headline, because the platform already saves the time. The MCP's value is the creative side: ideas that fit your audience, without browsing the full catalog. It suggests products and pre-made kits, and calculates prices as it goes.
Designs
You can select products and request designs from the chat. The design request goes into Sunday like any other. Sign-off against your brand guidelines stays with a person. More on the creative workflow in AI merch design.
Shipments, gifts and reporting
Plan a shipment, send merch out to people, send a gift, or request a report. Each shipment needs a human confirmation before it goes. Reporting comes back in the chat, where you can keep asking follow-up questions instead of building a new dashboard.
| What you ask | What Sunday MCP does | Who signs off |
|---|---|---|
| Stock check | Reads current stock by product and size | Nobody, it's a question |
| Data question | Counts items shipped, ranks your top countries | Nobody |
| Reorder | Builds a quote from what you ordered before | You, through the quote link |
| Budget split | Searches the catalog, proposes a split, creates a shareable wish list | You and your colleagues |
| Kit or product ideas | Suggests products and pre-made kits for your audience, calculates prices | You, when you request a quote |
| Design request | Selects products and requests designs | A person checks brand guidelines |
| Shipment or gift | Plans the shipment and prepares the send | You confirm the shipment |
| Reporting | Pulls the report into the chat | Nobody |
Thirty more, grouped by job, are in 30 prompts to run your merch program from Claude or ChatGPT.
Who approves what?
Sunday MCP cannot place an order. It generates a quote, shows it in the chat or sends it, and a person approves it through a link. Shipments need a human confirmation too. And design sign-off against your brand guidelines stays with a person.
Niels is blunt about it: "We will not process any orders without human verification." There are no automatic orders today.
The reason is the product itself. Merch is complex. There are designs, approvals and brand guidelines, and the output is physical. A wrong colour on a production run can't be fixed with a page refresh. So we automate as much of the preparation as possible, and a human confirms that the brand is respected before money moves.
The connection follows the same logic. You sign in with your own Sunday account through OAuth, so the assistant works with what your account can see and do. The full picture, including how this compares to general MCP security advice, is in spend controls for AI agents. If you're the person in finance or procurement who has to sign this off, the agentic commerce FAQ for procurement and finance is written for you.
Who is Sunday MCP for?
We used it internally first. The goal was simple: find what took our own teams a lot of time, and see what could give people that time back. Once that worked, we opened it up in two directions.
Partners and agencies
Agencies and partners get merch requests from their clients all the time. Say you run an agency and a client asks for ideas for a team offsite on a fixed budget. Instead of building a shortlist by hand, you ask the assistant for options that fit the audience, adjust them in the chat, and request a quote you can forward. They find everything much faster.
Power users with large merch programmes
Some companies run merch like a supply chain: many products, many regions, stock in a warehouse, a steady rhythm of reorders. Those teams wanted to test it early, and they're the ones who get the most from data questions and one-prompt reorders.
Who else tends to get value from it:
- Marketing teams planning event merch and campaign kits.
- HR and people teams keeping onboarding kits stocked. The HR side is covered in AI agents for HR and new hire welcome kits.
- Sales teams sending gifts to accounts at the right moment.
- Office and ops managers who own the swag cupboard and the company swag store.
Run your merch from the assistant you already use
Request access and we'll send you the connector URL for Claude, ChatGPT, Gemini or any MCP client.
Request Sunday MCP accessIncluded on every Sunday plan · Works with Claude, ChatGPT and Gemini · You approve every order
Physical products in any software layer
Sunday on its own in a chat window is useful. Sunday next to the other connectors in your assistant is where it gets interesting. Your CRM knows who matters. Your calendar knows when the events are. Your HR system knows who starts on Monday. None of them can put a hoodie on someone's desk.
Here are three scenarios Niels uses to explain it. They show how the pieces fit, they're not results from a named customer. Which other servers are worth connecting is covered in best MCP servers for marketing, HR and ops teams.
From a CRM signal to a gift
Say you've connected HubSpot and Sunday to the same assistant. You ask HubSpot which customers are best to send a gift to right now, the ones where a gift would have the most impact. Then you ask Sunday MCP what to gift them. It generates the packs in the Sunday app. You review them, approve, and only then do they ship.
You, in Claude: From HubSpot, pick the customers where a gift would have the most impact this month. Then use Sunday to suggest a pack for each and prepare it for my approval.
Sunday MCP: I've prepared a pack for each customer on that list, built from products in your catalog that fit their profile. They're ready in Sunday for you to review. None of them will ship until you confirm.
The full sales version of this workflow is in AI agents for sales, and the gifting basics in client onboarding gifts.
The end of the chain is still a real parcel in a real letterbox. The assistant prepares it, a person confirms the shipment, Sunday delivers it.
An event forecast with quotes attached
Say your team hands out between 1,000 and 1,500 items per event. You load in your event calendar and ask for a forecast for next quarter, plus quotes for the next three or six months. The assistant does the arithmetic. You decide which quotes to approve. Recurring versions of this are in AI workflow automation for swag programs, and the product side in event merchandise.
Leftover stock, matched to people
Every merch programme ends up with leftovers. With other systems connected, you can ask: "This is leftover stock in our Sunday platform. Which customers or employees should I reward with it?" The assistant matches stock to people. You decide whether to send. More scenarios like this are in 12 agentic AI examples in merch and gifting.
Leftover stock becomes a reward when it reaches the right person with a personal note. Matching stock to people is a question an assistant can answer in seconds.
None of these is a new feature. Each one is a question across two systems that nobody had the time to ask.
Which AI assistants does Sunday MCP work with?
Claude, ChatGPT and Gemini, and any other AI model or platform that supports MCP. In Niels's words: "Whatever tool you use, it connects." That matters because most companies don't pick one assistant. Marketing might live in ChatGPT, engineering in Claude, and the rest of the company in whatever came with their workspace.
The setup is the same everywhere. You add Sunday as a custom connector using the URL you receive, then sign in with your Sunday account. The step-by-step is in how to connect Sunday to Claude, ChatGPT or Gemini. For questions like which data the assistant can see, or how MCP compares to Claude skills, see the Sunday MCP FAQ.
Searching for "Claude connectors" or "Claude integrations" for merch? This is it. Sunday is added as a custom connector, and the same connector works for teams that order merch from ChatGPT or Gemini.
How is Sunday MCP different from gifting MCPs?
Two gifting platforms launched MCP servers in June 2026. Goody launched on 25 June, self-serve, with curated third-party gifts from 600+ brands, review before each gift sends and daily budget limits. Sendoso launched on 24 June as a request-based beta covering gifting and direct mail. Both are good tools if what you want is to choose and send gifts.
Sunday MCP is merchandise management. It works with your own branded products: designed, produced, warehoused and shipped through Sunday. Gifting is one part of it. "Send these three customers a nice gift" works, it gets generated and then a human approves it. But you can also check the stock of your own hoodie, request a new design or reorder your onboarding kit.
| Gifting MCPs | Sunday MCP | |
|---|---|---|
| What gets sent | Gifts and direct mail, including third-party products | Your own branded merch, plus gifts |
| Stock in your warehouse | Not the focus | Ask about it directly |
| Designs in your brand | Not the focus | Request designs, human brand sign-off |
| Reorders | Not the focus | One prompt, delivered as a quote |
The fair, side-by-side version is in Sunday MCP vs Goody and Sendoso. For the platform comparison outside of AI, see Sunday vs Sendoso.
How do you get access to Sunday MCP?
Access is on request, and it's included on every Sunday plan. Four steps:
- Request access through the form on the Sunday MCP page.
- Sunday grants access and you receive the connector URL.
- Add Sunday as a custom connector in Claude, ChatGPT, Gemini or another MCP client using that URL, then sign in with your Sunday account.
- Start with a read-only question, like your current stock, before you ask for your first quote.
No Sunday account yet? Create a free one first, then request access. If you want to see what the platform does on its own, how it works covers it.
Where this is heading
Niels's bet: around 20% of merch, maybe more, will be ordered by agents in about three years. Not most of it. A significant share, concentrated in reorders.
His reasoning is a single line. "New products is a creative project. Reordering is an administrative transaction." Administrative transactions are exactly what agents are good at, so for AI-powered companies the reorder will move to the assistant quickly. The creative projects will stay collaborative, with the assistant suggesting ideas and people deciding what represents the brand.
That split, and what it means for procurement, budgets and brand control, is the subject of our companion guide: agentic commerce for branded merch, what AI agents should order and what they shouldn't.
