Connect products.ai to your AI with MCP

Preview: products.ai in your AI opens by request. This page describes how the connection works at launch.

You can get Pius Bot's picks inside Claude, ChatGPT, Gemini or any other app that supports MCP. You request access, your MCP URL comes with it, and you add that URL to your AI app. After that, when you ask your AI a shopping question, it can ask products.ai for the pick, the reasons and where to buy it.

MCP (Model Context Protocol) is an open standard for connecting AI assistants to outside tools and data. The specification is public at modelcontextprotocol.io.

What do you get in your AI?

You get the same research products.ai chat does, delivered to your AI as data:

One pick that fits what you asked, with the reasons it fits.
A one-line why not for each alternative.
Where to buy: a shop that has the product in stock and delivers to you, with its price and the date it listed it. Shops in your country come first.
The exact product, matched on its barcode (GTIN), so your AI doesn't confuse the Sony WH-1000XM4 with the XM5.
Flagged links. Affiliate links are added after the pick, and every one is marked.

Your AI still writes the reply itself. It decides when to call products.ai and turns what comes back into an answer in its own words. So keep one habit: a price your AI quotes should come with a shop and a date. If it doesn't, ask your AI where the price came from.

How does access work?

Access is by request, and a person in Amsterdam reads every request.

Open the access form on our home page.
Fill in your work email, your company or project, and what you're interested in. The field for what you're building is optional, but fill it in anyway. A single concrete line tells the reader far more than an empty box.
Select the button. It opens your email app with a message to [email protected] already filled in. Nothing is sent until you press send. You can also write to [email protected] directly.
Your MCP URL comes with access.

Access in your AI has its own agreement, separate from the consumer Terms, and pricing is by request. Plans and billing shows how it sits next to products.ai chat.

What to expect on each platform

Setup differs per platform, so we don't publish click-by-click steps for other companies' apps. Their menus change between versions, and a guide that is wrong by next month helps nobody. In general, you'll meet one of these routes:

Your situation	What you do
Your app accepts an MCP URL	Paste your MCP URL where the app lets you add a connector or an MCP server.
Your platform offers a products.ai extension or connector	Install it.
You use developer tools with a config file	Add the URL to the tool's MCP configuration.
You're building your own product	Use the API or MCP under a business agreement instead.

The connection works in any MCP client, and that includes Claude, ChatGPT and Gemini. Some apps only let you add your own connectors on certain plans, or leave that choice to a workspace admin. If you can't find the option, check your app's help pages, and name the app you use when you request access.

What happens after you connect?

Ask your AI a shopping question the way you normally would. Take "noise-cancelling headphones under €300 for travel". With products.ai connected, your AI can come back with:

the Sony WH-1000XM5 as the best fit for flights;
a why not for the Sennheiser Momentum 4: longest battery, but less quiet on a plane;
bol.com as the place to buy it: €226,99, in stock, delivered in 1–2 days.

The wording depends on your AI. The pick, the shop and the reasons come from products.ai.

Where to go next

MCP: the details of the connection, including the tools your AI can call.
Affiliate links: how flagged links work and why they never change the pick.
Get started: products.ai chat, if you'd rather ask Pius Bot directly.