onx, powered by mcp

Connect any AI agent to your Order Management System

onX is the open, MCP-based protocol that lets AI agents query inventory, products, and fulfillment data in real time. No proprietary APIs, no custom integrations. Just a standardized interface between any compatible agent and fulfillmenttools.

What agents can do

Real fulfillment data, accessible through natural language

Through the fulfillmenttools MCP Server, AI agents query live operational data using plain language. No custom queries, no proprietary API calls. The agent asks, onX translates, and fulfillmenttools responds in real time.

Inventory queries

Check stock levels and availability across every facility in your network. Ask for a specific SKU or filter by location, all in real time.

Product catalog access

Retrieve product information and metadata on demand. Agents pull the records they need without touching the underlying API.

Variant lookup

Query product variants and their attributes directly. From size and color to SKU-level detail, available through a single request.

Agentic workflows

Feed live fulfillment data into automated decision-making. Agents access the inventory and product information they need to act with confidence.

Who it's for

Built for the teams that run commerce operations

The fulfillmenttools MCP Server gives different teams access to the same live operational data, each through the lens of their own work.

fulfillmenttools icon

Inventory managers

Query stock levels and availability across every facility in your network, without opening a dashboard.

fulfillmenttools icon

Merchandising teams

Access product catalog and variant information on demand to keep listings accurate and complete.

eCommerce platforms

Integrate live inventory and product data into automated workflows across your commerce stack.

Developers

Build agentic commerce applications on top of fulfillmenttools data, using an open standard.

The technical view

How onX works

The fulfillmenttools MCP Server acts as a protocol adapter between AI agents and the fulfillmenttools backend. It does not implement business logic, infer missing data, or make assumptions. It receives a request, validates it against the onX schema, and passes it through to fulfillmenttools in real time.

When a user asks an agent a question in natural language, the agent interprets the intent and invokes the right tool with structured parameters. onX translates that into the fulfillmenttools API format, the backend processes it against live production data, and the result flows back through the MCP Server. The agent then presents the answer in plain language.

Flowchart showing user request passing through AI Agent, Auth, MCP Server, backend, and returning response to user.

Security and compliance

Built for enterprise trust from the ground up

Agentic access to operational data only works if it is secure by design. Every request through the fulfillmenttools MCP Server runs on authenticated, permission-scoped access, with built-in human oversight.

Authenticated access

Every request uses OAuth 2.1 with PKCE. The agent always acts in the name of a user, never on its own, and it disallows implicit flows.

fulfillmenttools icon

Permission-scoped

Access is limited to what the individual user is allowed to see. The current release is read-only, with no write operations exposed.

Human in the loop

fulfillmenttools performs no autonomous high-impact actions. Users stay the decision-makers, and the agent asks for clarification rather than guessing.

fulfillmenttools icon

Data residency in Europe

The fulfillmenttools MCP Server runs in the same region as your tenant, on Google Cloud infrastructure hosted in Europe. Your data stays close to home.

Certified

fulfillmenttools is GDPR and ISO-certified, bringing enterprise-grade compliance to every interaction through onX.

Transparent by design

Backend errors are returned exactly as received, without modification. The fulfillmenttools MCP Server never infers, auto-fills, or bypasses validation rules.

Get started

Ready to connect your agents to fulfillmenttools?

See the fulfillmenttools MCP Server in action and find out how onX brings live fulfillment data to any compatible AI agent. Book a demo and we will walk you through it.

Frequently asked questions

Everything you need to know about the evolution of Agentic Order Management as well as how our agents work and connect with your platform.

All topics

Connect AI

Setup and integration

Connect AI

How do external AI agents connect to fulfillmenttools?

External AI agents connect through the fulfillmenttools MCP Server from any MCP-compatible client. It offers two endpoints: onX, which implements the open onX protocol for standardized agent-to-OMS communication, and the fulfillmenttools endpoint, which gives agents read access to the fulfillmenttools API through two tools, one to find the right operation and one to run it. You connect once per tenant, without a custom integration.

Connect AI

What is an MCP server?

An MCP (Model Context Protocol) server is a standardized interface that gives AI agents access to a system's data and functions. Instead of building a custom API integration for every AI tool, you connect once, and any MCP-compatible agent can use it. fulfillmenttools provides this access through the fulfillmenttools MCP Server, which supports onX, an open protocol built on MCP, and gives agents read access to the fulfillmenttools API.

Connect AI

What is onX?

Order Network eXchange (onX) is the open protocol for agent-to-OMS communication, built on the Model Context Protocol (MCP) and established by the Commerce Operations Foundation, of which fulfillmenttools is a founding member. Through onX, AI agents access real-time inventory, delivery promises and fulfillment data in the fulfillmenttools order management system, without custom integrations.

Connect AI

Which data can external AI agents access?

Through the fulfillmenttools MCP Server, external AI agents can access real-time inventory availability, product data, delivery promises, fulfillment options, and live order data, as well as facilities, pick and pack jobs, shipments, returns, and routing decisions. Access is limited to what the authenticated user is allowed to see, and the current release is read-only.

Setup and integration

Do I need to replace my existing systems to use AI in fulfillmenttools?

No. fulfillmenttools is API-first and follows MACH-certified, designed to integrate with your existing commerce stack without disrupting it. External AI agents connect via onX alongside your current systems. Internal agents operate within fulfillmenttools and can be activated progressively as your operations are ready for them.

Connect AI

Is AI agent access to fulfillment data secure?

Yes. Every request to the fulfillmenttools MCP Server uses OAuth 2.1 with PKCE, and the agent always requests data in the name of an authenticated user, never on its own. Access is permission-scoped to what that user is allowed to see, the current release is read-only, and data stays hosted on Google Cloud infrastructure in Europe.

Setup and integration

How quickly can we get started?

fulfillmenttools is built for fast implementation. Enterprise retailers are typically up and running in 9 weeks depending on their setup. Our team works with you from day one to configure fulfillmenttools, connect your systems, and activate the right level of AI for your current operational maturity.

‍