Give the assistant a price list, a spec sheet or a spreadsheet — it creates the materials, builds the configurator, runs test calculations and adjusts the formulas until the numbers match yours. You review the result instead of clicking through the editor.
MCP (Model Context Protocol) is an open standard that lets AI assistants use external tools. Configo's MCP server exposes your projects, materials and configurators to the assistant you already use — any MCP-compatible client. It works as you, with exactly your permissions, only after you approve it on a consent screen.
A price list, a spec, an Excel file or a plain description of the product — the assistant turns it into materials, controls, calculations and parts in the right format.
Runs calculations for the sizes and options you name and shows the price and bill of materials, so a wrong formula is caught before the first order.
"Add a markup for the premium fabric", "round prices to tens", "the hinge count should depend on height" — the configurator gets edited and the numbers re-checked.
Reuses materials you already have, creates only what's missing, updates prices from a new price list and warns when a unit or a price tier looks off.
In your assistant's connectors, add https://mcp.configo.org/mcp. The assistant finds it as a remote MCP server and registers itself.
Configo shows the usual login and consent screen. You grant access to projects, materials and configurators — nothing else.
"Build a configurator for Plisse mosquito screens from this price list" is enough to start. The assistant reads the schema reference itself.
Build a configurator from this price list and test it on three typical orders
Check every calculation in "Venetian blinds" and tell me which ones return zero
Add a material "Oak veneer, 2 mm" at 34 per m² and use it in the door configurator
Explain how the price of a 1600×2000 wardrobe is made up
Update material prices from this Excel file
Rename the product template so it includes the fabric and the width
The assistant sees and changes only what you could yourself, in the projects you belong to. There is no separate robot account with wider access.
The connection gets three scopes — projects, materials, configurators. Orders, contacts, billing and team settings stay out of reach.
The connection appears under Settings → Security like any other app. Revoke it there and it stops working immediately.
Model Context Protocol is an open standard for connecting AI assistants to tools and data. An MCP server describes what it can do; the assistant decides when to call it. Configo's server exposes configurator-building tools — the assistant uses them the same way it uses a calculator or a web search.
Any MCP-compatible client with remote server support: Claude (web, desktop and Claude Code), Cursor, and others. Most of them connect with Register automatically — no keys to copy. Clients that ask for a client ID can use a public OAuth application you create in Configo, see the docs.
Only what the assistant requests through the tools during your conversation: the projects, materials and configurators you're working on and the calculation results. Configo never sends your data to any AI provider on its own — the connection is initiated and controlled by you from your assistant.
It edits exactly what you ask for, and every change goes through the same validation as the editor. Ask it to test the numbers before you rely on them — that's what the calculation tool is for. Configurators keep working for orders and the widget while you iterate.
No. The MCP server is part of Configo on every plan. The assistant itself is your own subscription with its provider.
Sign up, connect your assistant and build your first configurator from a price list today.