// Tools
MCP Server.
MCP Server connects your lab instruments to any LLM that speaks the Model Context Protocol (MCP). Ask your lab a question in one sentence and get the answer. We set it up on your gateway, on request, and configure it to your devices and workflows. It runs locally, in your lab.
Any LLM that speaks MCP · Hosted or local LLM · Runs on your gateway · Per-device permissions
// Where it runs
Where MCP Server runs in your lab
MCP Server reaches your instruments through OPC UA / LADS. It runs on your gateway, in the lab you already operate. On request, we also deploy it on a server in your lab network. MCP Server needs network access to your instruments.
See the example prompts and the data flow on our AI and MCP page →
// On your gateway
MCP Server on Gateway 2 or Gateway Touch
If you run Gateway 2 or Gateway Touch, MCP Server runs as a component on the gateway. The gateway’s connectors provide the data path, and MCP Server provides the LLM interface.
This is the simplest route when your lab already uses SynConnect.
// Existing LADS servers
MCP Server for your existing LADS servers
Your lab already runs LADS servers or several gateways? On request, we deploy MCP Server in your lab and connect it through a Local Discovery Service (LDS) or another aggregation system. Tell us about your setup, and we confirm what’s possible for your instruments.
// Your LLM
Hosted or local: the LLM is your choice
MCP Server works with any LLM that speaks MCP:
- Hosted assistants: the MCP-capable assistant you already use. We set up the connection to the LLM you choose during deployment.
- Local LLMs: a model on your own hardware, so the whole loop stays inside your network
You choose, based on your data-handling preferences, your compliance requirements and your existing LLM contracts. With a hosted assistant, your prompts and the data MCP Server returns pass through that provider, under your contract. See our AI and MCP page for the data flow of each choice.
// Prerequisites
What you need to get started
- Gateway 2 or Gateway Touch to run on. For existing LADS servers, LDS or other aggregation systems, ask us.
- Network access to your instruments’ OPC UA ports, when MCP Server runs on a server in your lab network
- An LLM account or a local LLM, chosen and set up by you. For testing, you can also use a local LLM installation.
- A scoping call with us to define the action allow-list, the parameter scope and the permissions per device
SmartCable (coming Q1 2027) is a connectivity accessory that connects one instrument to a gateway or a PC. Ask us about the instruments you want to connect with it.
// How you buy
We set up MCP Server for your lab.
Our setup includes:
- Deployment on your gateway
- Connection to your data sources: the gateway’s connectors
- Action allow-list: the actions the LLM can call, for example which tools and skills are available
- Parameter scope: the data the LLM can read or write
- Permissions per device, from read-only access to full instrument control
We deliver MCP Server as a project with our team and quote it for your lab.
Want to see MCP Server with your instruments?
Tell us which instruments you want to connect, which LLM you use (hosted or local) and which questions your team asks. We show you a demo and confirm what’s possible for your instruments.
// Read more
Related pages
Integration
Model Context Protocol
How the Model Context Protocol works and how we use it in your integration.
Product
Gateway 2
MCP Server runs on Gateway 2, and every AI-driven action is written to the gateway’s audit trail. See how the Gateway protects your data.