Readying Queek so AI assistants can sell from your real catalog,through the Model Context Protocol.
MCP is the emerging standard that lets AI assistants use real tools and data instead of guessing from a web page. We are building toward MCP-compatible pathways so an assistant can discover a Queek OS catalog, check live pricing, and place a real order. A direction, not a finished spec.
AI assistant
Where the customer asks to buy
MCP
Tools and data, exposed as a contract
Queek OS
Real catalog, checkout, payments, delivery
Conceptual tools · the intended shape
- search_catalog
- check_availability
- create_cart
- place_order
- track_order
An AI assistant can only sell what it can actually reach. We are readying Queek OS to be reachable through MCP.
A standard way for AI assistants to use outside tools and data.
The Model Context Protocol is an open standard that lets an AI assistant call defined tools and read structured data from a system, rather than scraping a page and guessing. For commerce, that is the gap between an assistant that can describe a product and one that can check real stock, build a cart, and place an order.
- 01
It is a contract, not a scrape
Tools and data are exposed deliberately, with shapes an assistant can rely on.
- 02
It is why commerce gets real
Structured catalog, live availability, and a checkout an assistant can actually complete.
- 03
It is still emerging
MCP is young and moving. We treat it as a direction to build toward, not a settled spec.
What an AI assistant could do through Queek, over MCP.
Each capability below maps to data and operations Queek OS already runs for live businesses. Over MCP, an assistant would reach them through the merchant's real catalog, with the merchant's controls intact.
Discover the catalog
Read products, variants, and details as structured data, not scraped text.
Check availability
See live stock and hours, so it never offers what is already gone.
Resolve pricing
Real prices, fees, and options settled before anything is confirmed.
Build a cart
Assemble items into a real Queek OS cart on the customer's behalf.
Place an order
Complete checkout and payment as a genuine, confirmed order.
Track it
Read live order and delivery status back to the customer.
- search_catalog
Find merchants and items that match a request.
- check_availability
Confirm live stock, hours, and delivery reach.
- create_cart
Assemble items into a real Queek OS cart.
- place_order
Run checkout and payment, returning a real order.
- track_order
Read live status for the customer's assistant.
The intended shape, not a documented live spec. Names and fields will change.
What these tools could look like, framed honestly.
The tools below are the intended shape of an MCP surface over Queek OS, not a documented live spec. Names, fields, and behaviour will change as the protocol and our implementation mature. We are sharing the direction so early partners can help shape it.
The merchant and customer stay in control
- The merchant defines what an assistant can see and do on their catalog.
- Payment and fulfillment run through the same confirmed Queek OS flow as any order.
- A human handoff is always a step away when something needs a real answer.
- New
Online store
2x Jollof rice, 1x drink
- New
WhatsApp
Paracetamol, vitamin C
- New
AI assistant · MCPNew channel
Lunch + medicine to the office
- New
POS / counter
Walk-in, card payment
An MCP order is just another order in the one queue.
Nothing about MCP adds a parallel system to run. An order that originates from an AI assistant lands in the same Queek OS queue as web, WhatsApp, and counter sales, with the same catalog, checkout, payments, and delivery behind it.
One catalog
The same products and live stock every channel reads from.
One order queue
An MCP order sits beside web, WhatsApp, and POS orders.
One fulfillment path
Picked, paid, dispatched, and tracked like any other order.
What is real now, what we are building toward, and what we will not claim.
We would rather be clear than impressive. MCP support is a direction we are building toward, not a documented, production product you can connect to today.
- The operations an MCP surface needs, live on Queek OS
- Structured catalog, stock, and pricing as one source of truth
- Checkout, payments, fulfillment, and live tracking
- MCP-compatible pathways onto Queek OS
- The tool shapes an assistant would call
- Early access, shaped with partners
- A documented, live MCP server you connect to today
- Published tool schemas, versions, or rate limits
- Production AI-assistant commerce as a shipped product
Building an AI assistant that should sell from real catalogs? Talk to us.
MCP support is taking shape, and early partners help decide what the tools look like and what ships first. Tell us what you are building and we will tell you exactly what is possible today.
Straight answers about MCP on Queek.
What is real, what is a direction, and what we will not pretend exists yet.
Can I connect an AI assistant to Queek over MCP today?
Not as a documented, self-serve product. The operations an MCP surface needs, catalog, stock, pricing, checkout, payments, and delivery, run today on Queek OS. The MCP pathway onto them is a direction we are building toward. Talk to us about early access rather than pointing you at a spec that is not public yet.
Are the tool names and schemas on this page real?
They are the intended shape, not documented live tools. Names like search_catalog or place_order illustrate the direction so partners can react to it. Fields, names, and behaviour will change as MCP and our implementation mature, and we will not present them as final until they are.
Does an MCP order behave differently from a normal order?
No. An order that originates from an AI assistant lands in the same Queek OS queue, on the same catalog, checkout, payments, and fulfillment as web, WhatsApp, and counter sales. It is just another order, with the merchant's controls intact.
How is this different from your Agentic Commerce page?
Agentic Commerce is the merchant view of being ready for AI-driven demand. This page is the developer and protocol view of MCP specifically: how an assistant would reach Queek OS, what the tools could look like, and where that work honestly stands.
What about UCP, APIs, and webhooks?
MCP is one pathway. The Queek API is in early access and UCP-compatible pathways are a parallel direction. Webhooks are on the roadmap. The developers hub lays out each surface and how finished it is, plainly stated.
The operations are real today. The MCP pathway is what we are building next.
Start from the operating system that is already live, and help shape how AI assistants reach it over MCP.