Skip to main content

Overview & Concepts

What is Koongo MCP? Run your Koongo account by chatting with an AI assistant like Claude, ChatGPT or Gemini - key terms and how to stay safe.

What is Koongo MCP?

Koongo MCP lets you run your Koongo account by chatting with an AI assistant - such as Claude, ChatGPT or Gemini. Instead of clicking through screens, you describe what you want ("add an Amazon integration", "map my colours to Google's list", "build a Google Shopping feed") and the assistant does it for you by talking to Koongo behind the scenes.

You stay in control: the assistant proposes and performs each step, shows you what it did, and anything that would publish products to a live channel is protected by a safety step (see below).

You don't need to be technical. These guides give you ready-made sentences you can copy, paste and adapt. If you can write an email, you can use Koongo MCP.

The words you'll see

A few Koongo terms come up again and again. Here is what they mean in plain language.

Term

What it means

Project

Your connected store inside Koongo. Everything happens inside a project. If you have only one, the assistant selects it automatically; if you have several, it will ask which one.

Feed

A product data file Koongo generates for a channel. You may also hear "channel" or "channel profile" - it is the same thing. The channel either downloads the file or Koongo sends it.

Marketplace

A sales channel where you actually sell and receive orders back - Amazon, eBay, Kaufland, Bol, Zalando… It has a step-by-step setup wizard and order synchronisation.

Ad

An advertising or price-comparison channel - Google Shopping, Meta, Microsoft… It works exactly like a marketplace, but has no order sync (ads don't send orders back).

Store (source) attribute

A field that comes from your store - brand, EAN, price, quantity, colour…

Output (channel) attribute

A field the channel requires in the feed. Mapping connects your store attribute to the channel's required attribute.

Category mapping

Matching your store categories to the channel's own category tree (its taxonomy).

Source / enrichment

Extra product values you add yourself (for example an AI-derived colour or material) so you can map or filter on them.

How to talk to the assistant

  • Write naturally, one goal at a time: "Create a Google Shopping feed for my store."

  • Let it ask questions. If something is ambiguous (which project, which channel), it will check with you.

  • Ask it to explain or to show you before it commits: "Before you activate anything, show me what you're about to do."

  • Work in small steps and confirm as you go - the task guides in this section give you the exact sentences.

You rarely need to name your project. When your account has a single project, the assistant uses it automatically.

Staying safe: MANUAL vs AUTO

Finishing an integration's Summary step activates it. In AUTO product mode that publishes every matching product to the live channel (which can cost budget). When you are testing, ask the assistant to set MANUAL product mode first, publish one product as a dry run, and only then release the rest.

A good habit: "Set this marketplace to MANUAL mode, publish just one product so I can check it, and don't activate everything yet."

Before you can start

  • A Koongo account with at least one project (your store connected and imported).

  • Access to the MCP closed beta and an AI client connected - see Installation & Connection.

Once connected, pick a task guide: Create an Integration, Attribute Mapping, Enrich Products with New Attributes, Category Mapping, or Create & Export Feeds.

Did this answer your question?