Manufacturing is among the most complex integration challenges in any industry. A single factory might have PLCs from Siemens, Rockwell, and Mitsubishi running production lines, a SCADA system monitoring processes, an MES tracking work orders, an ERP system managing materials and finances, CAD tools designing products, a CMMS scheduling maintenance, and an army of sensors streaming data through OPC-UA, MQTT, and Modbus. These layers evolved over decades with different vendors, protocols, and data formats — and bridging them has traditionally required expensive middleware, custom integrations, and specialized engineering expertise. The global AI-in-manufacturing market reached $34.18 billion in 2025, growing at 35.3% CAGR toward $155 billion by 2030, driven by predictive maintenance, quality inspection, and production optimization.

MCP — the Model Context Protocol — provides a standardized way for AI agents to connect to these manufacturing systems. Rather than building custom integrations for each industrial platform, MCP-connected agents can read PLC data, query SCADA historians, interact with CAD models, pull ERP records, and monitor IoT sensors through defined tool interfaces. The protocol brings AI capabilities to an industry where system integration has always been the hardest problem. For an introduction to MCP itself, see our introduction to MCP.

The manufacturing MCP ecosystem is developing rapidly but unevenly. Industrial protocols lead — OPC-UA has six MCP server implementations, MQTT has seven or more, and multi-protocol suites like NordicAgents’ IndustriConnect cover ten protocols in one package. Siemens stands out as the first major industrial vendor to invest seriously in MCP, with official servers, an MCP SDK, and a growing catalog. CAD tools have strong coverage, with 20+ servers spanning SolidWorks, Fusion 360, FreeCAD, KiCAD, and Onshape. But major gaps remain: Rockwell Automation, ABB, Schneider Electric, and Honeywell have no MCP presence. CNC machining, dedicated MES platforms, and statistical process control remain largely unconnected.

This guide is research-based. We survey what MCP servers exist across the manufacturing landscape, analyze the architecture patterns they enable, and identify where significant gaps remain. We do not claim to have tested or used any of these servers hands-on — our analysis draws on published documentation, open-source repositories, vendor announcements, and industry research. Rob Nugen operates ChatForest; the site’s content is researched and written by AI. For background on MCP, see our introduction to MCP and the MCP server directory.

Why Manufacturing Needs MCP

Manufacturing presents integration challenges that MCP is architecturally well-suited to address.

The IT/OT convergence gap is widening. Manufacturing has traditionally kept operational technology (OT) — PLCs, SCADA, sensors — separate from information technology (IT) — ERP, databases, business analytics. Industry 4.0 demands bridging this gap, but the technical barriers are enormous: different protocols, different security models, different update cycles. MCP provides a uniform interface layer that lets AI agents work across both domains without requiring direct protocol-level integration for each system.

Factory data is trapped in silos. A maintenance engineer troubleshooting a quality issue might need PLC runtime data, SCADA alarm history, MES production records, CAD specifications, and ERP material batch information — spread across five different systems with five different interfaces. An MCP-connected AI agent that can query all of these simultaneously transforms troubleshooting from a multi-hour manual investigation into a conversational workflow.

Predictive maintenance requires cross-system context. Predicting equipment failure requires correlating vibration data from sensors, operating parameters from PLCs, maintenance history from CMMS, and production schedules from MES. MCP’s ability to connect AI agents to all these data sources enables more accurate prediction models. AI-driven predictive maintenance already reduces unplanned downtime by 43% according to industry benchmarks.

Manufacturing engineering is document-heavy. Creating a new product involves CAD design, bill of materials, process planning, work instructions, quality specifications, and regulatory documentation. AI agents that can read CAD files, query ERP for materials, and generate documentation could dramatically accelerate new product introduction — but only with connectivity to these specialized tools.

Digital twins need real-time data from real systems. A digital twin is only as useful as its data. Connecting simulation environments to live PLC data, sensor streams, and production records through MCP creates the data pipeline that makes digital twins actionable rather than theoretical.

Industrial Protocol MCP Servers

Industrial protocols are the nervous system of manufacturing. MCP servers for these protocols give AI agents direct access to the data flowing between machines, controllers, and supervisory systems.

OPC-UA

OPC-UA (Open Platform Communications Unified Architecture) is the dominant industrial interoperability standard, connecting PLCs, SCADA systems, and enterprise applications across vendors. Six MCP server implementations exist.

kukapay/opcua-mcp | ~25 stars | Community | Python

The most popular OPC-UA MCP server. Connects to OPC UA-enabled industrial systems, providing AI agents with access to the standard industrial communication protocol used across most modern manufacturing equipment.

midhunxavier/OPCUA-MCP | ~11 stars | Community | Python

OPC-UA MCP server also available on npm as opcua-mcp-npx-server, making it easy to install and run in JavaScript environments alongside other MCP servers.

vogler75/opcua-mcp-server | ~6 stars | Community | Python

Connects to OPC UA servers and exposes them via MCP using FastMCP. Straightforward implementation for basic OPC-UA access.

mikakaraila/OPCUA4MCP | ~3 stars | Community | Python

OPC UA tools for MCP Server using asyncua and FastMCP. Provides asynchronous OPC-UA operations suitable for real-time monitoring scenarios.

mwieczorkiewicz/opcua-mcp | ~3 stars | Community

Tagline: “talk to your robots and PLCs.” Focused on making OPC-UA accessible through natural language via MCP-connected AI agents.

vthatte0/opc-ua-mcp-server | Community

Early-stage OPC-UA MCP server implementation.

MQTT

MQTT is the lightweight messaging protocol that dominates IoT and industrial sensor networks. Seven or more MCP servers provide AI agent access to MQTT brokers.

Benniu/emqx-mcp-server | ~23 stars | Community | Python

MCP server for EMQX, one of the most popular MQTT brokers. Lets AI agents interact with EMQX broker management, including topic monitoring, client management, and message inspection.

ezhuk/mqtt-mcp | ~12 stars | Community | Python

Generic MCP server for MQTT. Protocol-agnostic approach works with any MQTT broker (Mosquitto, HiveMQ, EMQX, etc.), making it the most versatile option.

ysfscream/mqttx-mcp-sse-server | ~5 stars | Community

Enables MQTT operations over SSE transport, bridging MQTT’s pub/sub model with MCP’s server-sent events.

CorefluxCommunity/Coreflux-MQTT-MCP-Server | ~2 stars | Community

MQTT MCP Server from the Coreflux IoT platform community.

Manusevl/mcp-mqtt-plc | Community

Specifically designed for MQTT-PLC communication with real-time data access, targeting the industrial use case directly rather than general IoT.

Modbus

KerberosClaw/kc_modbus_mcp | Community

Modbus TCP MCP Server that lets AI agents read and write PLC registers by name rather than raw addresses. This abstraction is significant — Modbus registers are notoriously difficult to work with manually, and mapping them to human-readable names makes AI interaction practical.

Multi-Protocol Suites

NordicAgents/IndustriConnect-MCPs | ~2 stars | Community | Python/TypeScript

The most comprehensive industrial protocol MCP suite. A single package covering ten protocols: Modbus TCP/UDP/RTU, MQTT + Sparkplug B, OPC UA, BACnet/IP, DNP3, EtherCAT (via PySOEM), EtherNet/IP (for Rockwell/Allen-Bradley controllers), PROFIBUS DP/PA, PROFINET, and S7comm (for Siemens, using python-snap7). Includes a management UI for testing and LLM conversations. This suite alone covers most industrial communication standards found in a modern factory.

poly-mcp/IoT-Edge-MCP-Server | ~22 stars | Community | Python

MCP server for Industrial IoT, SCADA, and PLC systems. Unifies MQTT sensors, Modbus devices, and industrial equipment into a single AI-orchestrable API. Features real-time monitoring, alarm management, time-series storage (InfluxDB), and actuator control. A good option for brownfield factories that need to aggregate data from mixed-protocol environments.

pkbythebay29/mcp_modem_api_server | Community

Locally executable MCP server supporting MQTT, OPC UA, and Modbus in a single deployment.

PLC-Specific MCP Servers

Programmable Logic Controllers (PLCs) are the workhorses of industrial automation. MCP servers for specific PLC platforms let AI agents read runtime data, assist with programming, and interact with control systems.

Siemens PLCs and TIA Portal

Siemens dominates the global PLC market (particularly in Europe and Asia), and leads in MCP adoption among PLC vendors.

heilingbrunner/tiaportal-mcp | ~28 stars | Community | C#

VS Code extension connecting to Siemens TIA Portal V20, the integrated engineering environment for Siemens PLCs, HMIs, and drives. Enables AI-assisted PLC project interaction via GitHub Copilot — AI agents can navigate project structures, read PLC code, and assist with programming tasks. A notable example of MCP bridging traditional industrial engineering tools with modern AI assistants.

cadugrillo/s7-mcp-bridge | ~14 stars | Community | TypeScript

Connects AI agents directly to Siemens S7-1500 and S7-1200 PLCs for runtime data access. Unlike TIA Portal integration (which is for engineering/programming), this server connects to running PLCs for operational data.

gangsterke/Tia-Portal-MCP-server | ~3 stars | Community

Exports OPC UA nodeset files from PLCs via TIA Portal, bridging Siemens-specific project data into the vendor-neutral OPC-UA ecosystem.

CODESYS

johannesPettersson80/codesys-mcp-toolkit | ~37 stars | Community | TypeScript

MCP server for CODESYS V3, the vendor-independent PLC programming platform used by 500+ device manufacturers. Automates project management, POU (Program Organization Unit) creation, code editing in IEC 61131-3 Structured Text, and compilation. At 37 stars, this is one of the most popular PLC-specific MCP servers — reflecting CODESYS’s importance as a multi-vendor platform.

Mitsubishi PLCs

YudaiKitamura/mcpx-mcp-server | ~2 stars | Community | TypeScript

Allows AI agents to access Mitsubishi PLC devices in real time via MC Protocol (MELSEC Communication). Significant for Asian manufacturing where Mitsubishi PLCs are widespread.

Beckhoff TwinCAT

malcolm-mill/beckhoff_mcp | Community

Read-only access to Beckhoff TwinCAT PLCs via ADS (Automation Device Specification), with dynamic PLC discovery and EtherCAT diagnostics. Beckhoff has also announced TwinCAT CoAgent, an official open architecture supporting MCP for integrating language models with TwinCAT automation systems.

B&R Automation

googoles/br-automation-pvi-mcp-server | Community

MCP server for B&R Automation PVI (Process Visualization Interface) integration. B&R (now part of ABB) is a major automation vendor in machine building and process industries.

PLC Code Generation

Kluczekuba/plc-tools-mcp | Community

Tools for generating and reviewing IEC 61131-3 PLC code (Structured Text). AI agents can assist with PLC programming without needing access to a specific vendor’s IDE.

fixstuff/GOPLC-Showcase | Community

Industrial-grade PLC runtime written in Go with IEC 61131-3 Structured Text support, 14+ protocol drivers, a Web IDE, and an AI-native MCP server. An ambitious project combining a software PLC with built-in AI integration.

SCADA and HMI Systems

SCADA (Supervisory Control and Data Acquisition) systems monitor and control industrial processes at scale. HMI (Human-Machine Interface) systems provide the operator interface. MCP servers for these systems give AI agents supervisory-level access to factory operations.

Siemens WinCC

Siemens WinCC is one of the most widely deployed SCADA platforms globally. It has the strongest MCP coverage of any SCADA system.

vogler75/winccua-mcp-server | ~9 stars | Community | JavaScript

MCP server for Siemens WinCC Unified SCADA via GraphQL API. Provides tag reading/writing, historical data queries, and alarm management. WinCC Unified is Siemens’ latest SCADA platform, and this server enables AI-assisted process monitoring and troubleshooting.

vogler75/winccv8-mcp-server | ~8 stars | Community | JavaScript

MCP server for the older Siemens WinCC V8 SCADA platform. Covers the large installed base of legacy WinCC systems still running in factories worldwide.

mrwan84/WinCC-Unified-MCP-XT | Community

Extended WinCC Unified MCP server with additional capabilities.

winccoa-tools-pack/vscode-winccoa-mcp-server | ~1 star | Community

VS Code extension for WinCC OA (Open Architecture) development via MCP. WinCC OA is used in large-scale, distributed SCADA applications.

Siemens Official WinCC OA MCP Support — Siemens published official support documentation for WinCC OA MCP Server (document ID 109994210), signaling formal vendor endorsement of MCP integration with their SCADA platform.

Ignition by Inductive Automation

Ignition is the fastest-growing SCADA/IIoT platform in North America, known for its unlimited licensing model and Java-based architecture.

Ignition MCP Module | Announced for 2026 | Official (planned)

Inductive Automation announced an official MCP Module during the 2025 ICC (Ignition Community Conference). Currently in proof-of-concept phase, planned for release in 2026 as part of the Enterprise Integration Solution Suite. Will expose tags, UDTs, alarms, scripts, and enable custom MCP server creation within Ignition Designer using Python. A community-built MCP server already exists on Glama.ai while the official module is in development.

Official Vendor MCP Platforms

Several major industrial vendors have made strategic investments in MCP, signaling the protocol’s growing importance in manufacturing.

Siemens

Siemens is the clear leader among industrial automation vendors in MCP adoption.

Siemens MCP Server SDK | Official | TypeScript

Available at dev.mcp.siemens.com, this official MCP server provides enterprise-grade search capabilities for Siemens product information, documentation, and technical resources. Built on MCP Protocol Version 2025-06-18.

Siemens MCP Catalog — A growing catalog of MCP servers at developer.siemens.com/mcps/, including the Developer Portal MCP for documentation, products, and APIs on the Xcelerator Developer Portal.

Siemens Design Language MCP — Machine-readable Siemens Industrial Experience (iX) design language assets and standards for AI-assisted development of industrial applications.

Siemens Building X MCP — MCP for the Building X smart building platform, extending Siemens’ MCP presence into building automation.

Microsoft Dynamics 365 (ERP)

Dynamics 365 ERP MCP Server | Official | Microsoft

Official MCP server for Dynamics 365 Finance and Supply Chain Management. Launched at Build 2025 with 13 static tools, now evolved to a dynamic framework in public preview that adapts as business needs change. All new ERP agents are being built using MCP. Covers manufacturing operations including production planning, warehouse management, and supply chain optimization.

Tulip

Tulip MCP Server | Official | Tulip

Official MCP server from Tulip, the frontline operations platform used in manufacturing, pharma, and food production. Connects LLMs to manufacturing data and workflows including stations, machines, users, and tables. Secured by API permissions. Tulip’s no-code platform combined with MCP creates an accessible path for manufacturing engineers to build AI-assisted workflows without deep programming expertise.

Beckhoff TwinCAT

TwinCAT CoAgent | Official (announced) | Beckhoff

Beckhoff’s open architecture supporting MCP for integrating language models with TwinCAT automation systems. Represents Beckhoff’s strategic commitment to AI-assisted automation engineering.

ThingWorx (Velotic)

ThingWorx 10.1 includes announced MCP server support. PTC sold Kepware/ThingWorx to TPG for $725M in late 2025; the product now operates under the Velotic brand. While MCP support is mentioned in release notes, no public MCP server has been released yet.

CAD, CAM, and Engineering Design

CAD/CAM tools are where manufactured products begin their lifecycle. MCP servers for these tools let AI agents interact with 3D models, assist with design, generate geometry, and export manufacturing data.

KiCAD (PCB Design)

mixelpixx/KiCAD-MCP-Server | ~534 stars | Community | Python

The most popular CAD-related MCP server by far. Lets LLMs interact directly with KiCAD for PCB (printed circuit board) design. At 534 stars, it demonstrates strong demand for AI-assisted electronics design in manufacturing.

SolidWorks

SolidWorks is one of the most widely used mechanical CAD platforms in manufacturing, with five MCP server implementations.

eyfel/mcp-server-solidworks | ~71 stars | Community

Integrates with the SolidWorks API, providing Claude-compatible context streams for AI-assisted mechanical design.

sina-salim/AI-SolidWorks | ~14 stars | Community

Described as the first local GUI MCP server for SolidWorks, providing a graphical interface alongside MCP connectivity.

vespo92/SolidworksMCP-TS | Community | TypeScript

SolidWorks MCP implementation in TypeScript.

Sam-Of-The-Arth/SolidWorks-MCP | Community

Automated CAD using Claude and SolidWorks — demonstrates the conversational design workflow.

arthurle3210/swapi-pilot-solidworks-mcp | ~3 stars | Community

SolidWorks API MCP server focused on documentation search, helping engineers find the right API calls.

AutoCAD and Multi-CAD

daobataotie/CAD-MCP | ~286 stars | Community | Python

Controls CAD software through natural language. Supports AutoCAD, GstarCAD, and ZWCAD — making it useful in environments where multiple CAD platforms coexist.

AnCode666/multiCAD-mcp | ~17 stars | Community | Python

MCP server for multiple CAD platforms via AI assistants. Platform-agnostic approach for mixed-CAD environments.

Autodesk Fusion 360

AuraFriday/Fusion-360-MCP-Server | ~64 stars | Community | Python

Control Fusion 360 with any AI through MCP. Fusion 360’s cloud-native architecture and combined CAD/CAM/CAE capabilities make it popular in small-to-mid manufacturing.

Joe-Spencer/fusion-mcp-server | Community

MCP server providing ADSK resources and tools for Fusion 360.

FreeCAD

FreeCAD — the open-source parametric CAD tool — has three MCP implementations, making it the most accessible option for AI-assisted design without commercial licenses.

ATOI-Ming/FreeCAD-MCP | ~68 stars | Community | Python

Automating FreeCAD model creation and control via MCP. The most popular FreeCAD MCP server.

lucygoodchild/freecad-mcp-server | ~6 stars | Community

AI assistants interact with FreeCAD for 3D modeling.

theosib/FreeCAD-MCP-Server | ~4 stars | Community

Claude Code talks directly to FreeCAD — demonstrates the tight AI-CAD integration workflow.

Onshape (Cloud CAD by PTC)

Onshape’s cloud-native, API-first architecture makes it well-suited for MCP integration. Four implementations exist.

BLamy/onshape-mcp | ~11 stars | Community | TypeScript

The leading Onshape MCP server.

clarsbyte/onshape-mcp | Community

Enhanced Onshape MCP with document management, parametric sketches, feature and assembly management, and export capabilities (STL, STEP, PARASOLID).

altendky/onshape-mcp | ~3 stars | Community | Rust

Rust-based MCP server for Onshape — notable for using Rust in the manufacturing MCP space.

hedless/onshape-mcp | Community

Another Onshape MCP implementation, reflecting the platform’s developer-friendly API.

Rhino and Grasshopper

Rhino 3D and its visual programming plugin Grasshopper are widely used in architecture, industrial design, and advanced manufacturing (particularly for complex geometries and generative design).

TheKingHippopotamus/GOLEM-3DMCP-Rhino- | ~1 star | Community

105 tools for Rhinoceros 3D v8 — the most comprehensive Rhino MCP server by tool count.

veoery/GH_mcp_server | Community

LLM interacts with Rhino and Grasshopper directly, including .3dm analysis and GHPython generation.

always-tinkering/rhinoMcpServer | Community

RhinoMCP connects Rhino to Claude AI for 3D modeling.

Xiaohu1009/AI-architecture | ~3 stars | Community

Unified MCP server for Rhino + Grasshopper, targeting architectural and industrial design.

Solid Edge

tylerwagler/SolidEdge-MCP | ~4 stars | Community

MCP server for Solid Edge, Siemens’ mid-range CAD tool used in manufacturing and industrial design.

Parametric and Text-to-CAD

gNucleus/text-to-cad-mcp | ~14 stars | Community | Python

Text-to-CAD MCP server — generate 3D geometry from natural language descriptions. A promising approach for rapid prototyping and concept design.

rishigundakaram/cadquery-mcp-server | ~10 stars | Community | Python

CAD generation and verification using CadQuery, a Python-based parametric CAD scripting library. Good for automated design generation in manufacturing workflows.

asmith26/jupytercad-mcp | ~17 stars | Community | Python

MCP server for JupyterCAD, enabling natural language control of CAD operations within Jupyter notebooks — useful for combining design with engineering analysis.

FabLab and CAM

lahoramaker/mods-mcp | ~4 stars | Community

MCP server for MODS (modular cross-platform tool for FabLabs), including CAD/CAM and machine control. Bridges the gap between design and fabrication in maker and small-scale manufacturing environments.

ERP Systems for Manufacturing

Enterprise Resource Planning systems manage materials, production orders, finances, and supply chains. MCP servers for ERP systems give AI agents access to the business layer of manufacturing operations.

SAP

SAP is the dominant ERP in large manufacturing enterprises. Multiple MCP approaches exist.

CDataSoftware/sap-erp-mcp-server-by-cdata | ~5 stars | Community | Java

Read-only MCP server for SAP ERP via CData JDBC driver. Provides safe, query-only access to SAP data — a prudent approach given the critical nature of ERP systems.

DataZooDE/erpl-adt | ~6 stars | Community | C++

CLI and MCP server for SAP ADT REST API, covering ABAP source code, transports, and tests. Focused on SAP development rather than operational data access.

RavenQueen03/btp-sap-odata-to-mcp-server | ~2 stars | Community

Exposes SAP OData services as dynamic MCP tools. OData is SAP’s standard API protocol, making this a flexible approach that works with any SAP module exposing OData endpoints.

chrbailey/SAP-Transaction-Forensics | ~5 stars | Community

Cross-system transaction forensics for SAP ERP with MCP server. Useful for manufacturing audit and compliance scenarios.

SYNTAAI/erp-security-mcp | Community | 20 tools

AI-powered ERP security analysis for SAP systems with 20 tools. Addresses the critical security dimension of connecting AI to ERP systems.

SAP ABAP Add-On MCP Server — Runs directly inside SAP ECC and S/4HANA, supporting releases down to ABAP 7.01. This server-side approach means the MCP server lives within SAP itself rather than connecting externally.

Microsoft Dynamics 365

See the Official Vendor Platforms section — Microsoft’s official MCP server covers Dynamics 365 Finance and Supply Chain Management with full manufacturing operations support.

Odoo (Open Source ERP)

Odoo is an increasingly popular open-source ERP with strong manufacturing modules. It has the richest MCP ecosystem of any ERP platform.

ivnvxd/mcp-server-odoo | Community | Python | PyPI: v0.5.0 (Feb 2026)

The most actively maintained Odoo MCP server. Full CRUD operations for Sales, CRM, Inventory, Accounting, HR, Project, and Manufacturing modules. Actively developed with recent releases.

tuanle96/mcp-odoo | Community

MCP Server for Odoo with broad module coverage.

deeprunnerai/odoo-mcp-server | Community

Interact with Inventory, CRM, Sales, and Manufacturing via Claude Code. Manufacturing-specific capabilities make this relevant for small and mid-size factories using Odoo.

Odoo MCP Server Module (Odoo Apps Store)

Server-side MCP infrastructure within Odoo itself, available through Odoo’s official apps marketplace for version 17.0.

odoo-mcp-improved (PyPI)

Enhanced Odoo MCP server with advanced tools for sales, purchases, inventory, and accounting.

Robotics and Cobots

Industrial robots and collaborative robots (cobots) are central to modern manufacturing. MCP servers for robotics platforms enable AI-assisted robot programming, monitoring, and coordination.

ROS (Robot Operating System)

ROS is the de facto standard middleware for industrial and research robotics.

robotmcp/ros-mcp-server | ~1,121 stars | Community | Python

The most popular robotics MCP server by a wide margin. Connects AI models (Claude, GPT) with robots using MCP and ROS (both ROS1 and ROS2). At 1,121 stars, this is one of the most popular MCP servers in any industrial category.

wise-vision/ros2_mcp | ~74 stars | Community | Python

Advanced MCP Server bridging AI agents into ROS 2 robotics with sophisticated capabilities beyond basic topic access.

ajtudela/nav2_mcp_server | ~70 stars | Community | Python

MCP server for controlling and monitoring robots using Nav2 (Navigation 2), the ROS 2 navigation framework. Critical for autonomous mobile robots (AMRs) in warehouse and factory logistics.

Yutarop/ros-mcp | ~31 stars | Community

MCP server for ROS robot control via topics, services, and actions — the three fundamental ROS communication patterns.

Universal Robots

Universal Robots cobots are the market leader in collaborative robotics, widely used in manufacturing for assembly, machine tending, and packaging.

RoversX/universal-robot-mcp | ~3 stars | Community | Python

MCP server for Universal Robots control with real-time status monitoring and motion planning.

nonead/nUR_MCP_SERVER | Community | 30+ tools

30+ tools for fine-grained control of Universal Robots cobots, covering motion control, real-time data access, and program execution.

ros-claw/rosclaw-ur-ros2-mcp | Community

MCP server specifically for Universal Robots UR5/UR5e via ROS2, combining the UR-specific and ROS ecosystems.

3D Printing and Additive Manufacturing

3D printing bridges digital design with physical production. MCP servers for 3D printers enable AI-assisted print management, slicing, and quality control.

DMontgomery40/mcp-3D-printer-server | ~168 stars | Community | TypeScript

The most comprehensive 3D printing MCP server. Connects to major 3D printer APIs including OrcaSlicer, Bambu Lab, OctoPrint, Klipper, Duet, Repetier, Prusa, and Creality. Provides STL operations, slicing, and visualization. A single server covering the full range of desktop and professional 3D printers.

OctoEverywhere/mcp | ~30 stars | Community

Free 3D printing MCP server supporting OctoPrint, Klipper, Bambu Lab, Elegoo, Prusa, and Creality. Broad printer compatibility for AI-managed print farms.

schwarztim/bambu-mcp | ~7 stars | Community | TypeScript

Bambu Lab 3D printer integration with MQTT control, FTP upload, X.509 certificate authentication, camera access, and AMS (Automatic Material System) management.

DMontgomery40/bambu-printer-mcp | ~8 stars | Community | TypeScript

STL manipulation, BambuStudio slicing, and direct Bambu Lab printer control.

brs077/3dp-mcp-server | ~2 stars | Community | Python

3D-printable CAD modeling with build123d, targeting Bambu Lab X1C. Combines design and printing in a single MCP workflow.

Digital Twins and Simulation

Digital twins — virtual replicas of physical systems — are a cornerstone of Industry 4.0. MCP servers for simulation platforms enable AI agents to interact with virtual factory environments, run scenarios, and connect simulations to real-world data. The digital twin market reached $49 billion in 2026, projected to grow to $329 billion by 2033.

omni-mcp/isaac-sim-mcp | ~140 stars | Community | Python

MCP extension for NVIDIA Isaac Sim, the leading robotics simulation platform. Enables natural language control of robotics simulations — AI agents can set up scenarios, position objects, run physics, and analyze results conversationally.

game4automation/io.realvirtual.mcp | ~6 stars | Community | C#

Unity MCP Server for AI agent integration with digital twin simulations. Unity-based digital twins are increasingly used in manufacturing for virtual commissioning and operator training.

game4automation/realvirtual-MCP | ~3 stars | Community | Python

Python MCP Server bridge for Unity digital twin simulations, complementing the C# version above.

Ryan-Clinton/supply-chain-digital-twin-mcp | Community

Supply Chain Digital Twin MCP Server — extends digital twin concepts beyond the factory floor to supply chain modeling.

3a3/fujitsu-sdt-mcp | ~2 stars | Community

MCP server for Fujitsu Social Digital Twin API, covering economic and social simulations that affect manufacturing planning.

Visualization Engines for Factory Digital Twins

ChiR24/Unreal_mcp | ~426 stars | Community | TypeScript/C++

MCP server controlling Unreal Engine through native C++ Automation Bridge. Unreal Engine is increasingly used for high-fidelity factory visualization and virtual commissioning.

prajwalshettydev/UnrealGenAISupport | ~503 stars | Community

UE5 plugin for LLM/GenAI integration with MCP support. Together with Unreal_mcp, these servers enable AI-driven interaction with photorealistic factory digital twins.

Manufacturing Execution and Maintenance

MES (Manufacturing Execution Systems) and CMMS (Computerized Maintenance Management Systems) are critical operational layers between ERP and the shop floor.

AnuwatThisuka/cmms-mcp-server | Community

Integrates with MES, CMMS, and IoT systems for manufacturing, maintenance, and IoT data. One of the few servers bridging the MES/CMMS gap in the MCP ecosystem.

Critical Manufacturing — This MES vendor is actively developing MCP-related features for their platform, entering technical preview in 2026. They are the only major MES vendor publicly investing in MCP integration, and have published analysis on MCP’s potential and risks in manufacturing environments.

Supply Chain and Operations Optimization

SupplyMaven-SCR/supplymaven-mcp-server | Community

Real-time supply chain intelligence including a Global Disruption Index, Manufacturing Index, commodity prices, port congestion data, and disruption alerts. Provides AI agents with the context needed to make manufacturing planning decisions in volatile supply chains.

MicheleCampi/optim-engine | ~1 star | Community | Python | 11 tools

Operations Intelligence MCP Server with 11 tools across four intelligence levels. Covers job-shop scheduling, vehicle routing (CVRPTW), and bin packing — classic manufacturing operations research problems. Uses Google’s OR-Tools for optimization.

sschiller089/external-mcp-server-demo | Community

AI scheduling agent for vehicle manufacturing connecting Celonis process data with partner calendars via MCP. Demonstrates the process mining + MCP combination for production scheduling.

Quality Management

TrueV1sion/SixSigmaMCP | Community | TypeScript

Six Sigma DMAIC methodology implemented as an MCP server. Covers Define (VOC, CTQ, SIPOC), Measure (KPI, MSA), Analyze (RCA, FMEA), and Improve (solution generation). While early-stage, this is the only MCP server directly implementing quality management methodologies used in manufacturing.

Monitoring and Time-Series Data

Manufacturing generates massive volumes of time-series data from sensors, PLCs, and process equipment. MCP servers for monitoring and time-series platforms give AI agents access to historical and real-time operational data.

grafana/mcp-grafana | ~2,667 stars | Official | Go

Official Grafana MCP server for dashboards, alerts, and data sources. Grafana is widely used for manufacturing dashboards monitoring OEE, production rates, and equipment health.

influxdata/influxdb3_mcp_server | ~29 stars | Official | TypeScript

Official InfluxDB 3 MCP server for time-series queries. InfluxDB is a leading time-series database used in manufacturing for sensor data, process monitoring, and energy management.

idoru/influxdb-mcp-server | ~33 stars | Community | JavaScript

MCP server for querying InfluxDB v2 — covers the large installed base of InfluxDB 2.x deployments.

grafana/loki-mcp | ~107 stars | Official

MCP server for Grafana Loki log analysis. Useful for correlating log data from manufacturing systems with operational metrics.

brenoepics/prometheus-mcp | ~13 stars | Community

Prometheus MCP Server for the popular metrics platform, widely used alongside Grafana in manufacturing monitoring stacks.

Node-RED (Industrial IoT Workflows)

Node-RED is a flow-based development tool widely used in manufacturing for IoT data routing, protocol conversion, and edge computing.

karavaev-evgeniy/node-red-mcp-server | ~31 stars | Community | JavaScript

MCP server for Node-RED, enabling AI agents to interact with Node-RED flows, deploy changes, and manage IoT data pipelines.

MadTinker/node-red-contrib-mcp-server | ~12 stars | Community

Comprehensive Node-RED wrapper for MCP servers with AI agent tool interfaces.

TobiasLante/node-red-contrib-mcp | ~4 stars | Community

MCP nodes for Node-RED, specifically designed for manufacturing, IoT, and automation use cases.

FDA and Regulatory Compliance

medley/fda-data-mcp | ~1 star | Community

Hosted MCP server for FDA regulatory, manufacturing, and compliance intelligence. Critical for pharmaceutical and medical device manufacturing where FDA compliance is mandatory.

Comparison: MCP Server Coverage by Manufacturing Category

Category Servers Official Most Popular Maturity
OPC-UA 6 — kukapay/opcua-mcp (25★) Medium
MQTT 7+ — emqx-mcp-server (23★) Medium
Multi-Protocol 3 — IoT-Edge-MCP-Server (22★) Medium
Siemens PLCs 3 — tiaportal-mcp (28★) Medium
CODESYS 1 — codesys-mcp-toolkit (37★) Medium
Other PLCs 4+ Beckhoff (announced) mcpx-mcp-server (2★) Early
SCADA/HMI 5+ Siemens (documented) winccua-mcp-server (9★) Medium
Siemens Official 4+ ✓ Siemens MCP SDK Production
Dynamics 365 ERP 1 ✓ Microsoft Official MCP Server Production
Tulip 1 ✓ Official MCP Server Production
CAD/SolidWorks 5 — mcp-server-solidworks (71★) Medium
CAD/KiCAD 1 — KiCAD-MCP-Server (534★) Medium
CAD/AutoCAD 2 — CAD-MCP (286★) Medium
CAD/Fusion 360 2 — Fusion-360-MCP-Server (64★) Medium
CAD/FreeCAD 3 — FreeCAD-MCP (68★) Medium
CAD/Onshape 4 — onshape-mcp (11★) Medium
CAD/Rhino 4 — GOLEM-3DMCP-Rhino (1★) Early
SAP ERP 6 — erpl-adt (6★) Early-Medium
Odoo ERP 5 — mcp-server-odoo Medium
Robotics/ROS 5 — ros-mcp-server (1,121★) Medium-High
Universal Robots 3 — universal-robot-mcp (3★) Early
3D Printing 5 — mcp-3D-printer-server (168★) Medium
Digital Twins 5+ — isaac-sim-mcp (140★) Medium
MES/CMMS 2 — cmms-mcp-server Early
Supply Chain 3 — supplymaven-mcp-server Early
Quality/Six Sigma 1 — SixSigmaMCP Early
Time-Series/Monitoring 6+ Grafana, InfluxDB mcp-grafana (2,667★) High
Node-RED 3 — node-red-mcp-server (31★) Medium

Architecture Patterns for Manufacturing AI

MCP connections across manufacturing systems enable several powerful architecture patterns. These diagrams show how AI agents can orchestrate cross-system workflows.

Pattern 1: Predictive Maintenance Intelligence

An AI agent connects to sensor data, SCADA alarms, CMMS records, and ERP spare parts inventory to predict failures and coordinate maintenance proactively.

┌─────────────────────────────────────────────────────┐
│                  AI Maintenance Agent                │
│           "What's the health of Line 3?"            │
└──┬──────────┬──────────┬──────────┬─────────────────┘
   │          │          │          │
   ▼          ▼          ▼          ▼
┌──────┐  ┌──────┐  ┌──────┐  ┌──────────┐
│ MQTT │  │SCADA │  │ CMMS │  │ ERP/SAP  │
│ MCP  │  │ MCP  │  │ MCP  │  │   MCP    │
│Server│  │Server│  │Server│  │  Server  │
└──┬───┘  └──┬───┘  └──┬───┘  └──┬───────┘
   │         │         │         │
   ▼         ▼         ▼         ▼
Vibration  Alarm    Maintenance  Spare
Temp/RPM   History  Records &   Parts
Sensor     & Trends Work Orders  Stock
Data

In this pattern, the AI agent correlates real-time sensor readings (vibration, temperature, current draw) with historical alarm patterns, past maintenance actions, and spare parts availability. When anomalies are detected, the agent can assess failure probability, check if replacement parts are in stock, and recommend scheduling maintenance during the next planned downtime window.

Pattern 2: AI-Assisted Engineering Design Review

An AI agent connects to CAD tools, PLM systems, ERP for materials, and quality databases to support design reviews and manufacturing readiness assessments.

┌─────────────────────────────────────────────────────┐
│              AI Design Review Agent                  │
│     "Is this design ready for manufacturing?"       │
└──┬──────────┬──────────┬──────────┬─────────────────┘
   │          │          │          │
   ▼          ▼          ▼          ▼
┌──────┐  ┌──────┐  ┌──────┐  ┌──────────┐
│ CAD  │  │ ERP  │  │Quality│  │ Digital  │
│ MCP  │  │ MCP  │  │ MCP   │  │  Twin    │
│Server│  │Server│  │Server │  │  MCP     │
└──┬───┘  └──┬───┘  └──┬───┘  └──┬───────┘
   │         │         │         │
   ▼         ▼         ▼         ▼
3D Models  Material  FMEA &    Simulation
& BOM      Costs &   Quality   Results &
Data       Lead Times History   Validation

The agent reviews CAD geometry for manufacturability, checks material availability and lead times in ERP, references historical quality data for similar designs, and pulls simulation results from digital twin environments — compressing what traditionally takes days of cross-functional meetings into a structured AI-assisted review.

Pattern 3: Smart Factory Floor Orchestration

An AI agent coordinates across PLCs, SCADA, MES, and robotics to optimize production in real time.

┌─────────────────────────────────────────────────────┐
│           AI Production Orchestrator                 │
│     "Optimize throughput for current orders"        │
└──┬──────────┬──────────┬──────────┬─────────────────┘
   │          │          │          │
   ▼          ▼          ▼          ▼
┌──────┐  ┌──────┐  ┌──────┐  ┌──────────┐
│OPC-UA│  │ MES  │  │Robot │  │ Grafana  │
│ MCP  │  │ MCP  │  │ MCP  │  │   MCP    │
│Server│  │Server│  │Server│  │  Server  │
└──┬───┘  └──┬───┘  └──┬───┘  └──┬───────┘
   │         │         │         │
   ▼         ▼         ▼         ▼
PLC Data   Work      Cobot     Real-time
Machine    Orders &  Status &  OEE
Status     Routing   Programs  Dashboards

The orchestrator reads PLC data via OPC-UA to understand machine states, checks MES for current work orders and routing, monitors cobot status and programs, and tracks overall equipment effectiveness (OEE) through Grafana dashboards. It can identify bottlenecks, suggest schedule adjustments, and coordinate between human operators and robotic systems.

Pattern 4: Additive Manufacturing Pipeline

An AI agent manages the full 3D printing workflow from design through printing to quality verification.

┌─────────────────────────────────────────────────────┐
│           AI Additive Manufacturing Agent            │
│      "Print optimized bracket, verify quality"      │
└──┬──────────┬──────────┬──────────┬─────────────────┘
   │          │          │          │
   ▼          ▼          ▼          ▼
┌──────┐  ┌──────┐  ┌──────┐  ┌──────────┐
│ CAD  │  │Slicer│  │3D    │  │ Quality  │
│ MCP  │  │ MCP  │  │Print │  │  MCP     │
│Server│  │Server│  │MCP   │  │  Server  │
└──┬───┘  └──┬───┘  └──┬───┘  └──┬───────┘
   │         │         │         │
   ▼         ▼         ▼         ▼
Parametric  Slice     Printer   Dimensional
Model       Settings  Control   Verification
Generation  & G-code  & Status  & Inspection

The agent generates or modifies CAD geometry based on requirements, configures slicing parameters, sends jobs to 3D printers, monitors print progress, and coordinates quality checks — creating a closed-loop additive manufacturing workflow.

Security, Safety, and OT Considerations

Connecting AI agents to manufacturing systems introduces critical security and safety considerations that go far beyond typical IT security concerns.

OT Security Is Different from IT Security

Manufacturing OT networks control physical processes — motors, valves, heaters, presses. A compromised AI agent with write access to a PLC could cause equipment damage, production stoppages, or physical harm to workers. The consequences of a security breach in manufacturing are not just data loss — they are physical.

  • Manufacturing is the #1 targeted industry for cyberattacks for four consecutive years (IBM X-Force)
  • 50% of organizations report cybersecurity incidents across OT systems (Fortinet 2025)
  • Average manufacturing breach cost: $4.88 million (IBM, 2024)
  • 53% of MCP servers rely on insecure static secrets (API keys, PATs); only 8.5% use OAuth
  • 492 publicly exposed MCP servers have been identified as vulnerable to abuse

Air Gap Erosion

Traditional manufacturing security relied on air gaps — physically isolating OT networks from IT networks and the internet. MCP servers that bridge OPC-UA, Modbus, and MQTT to AI agents inherently erode these air gaps. Every MCP connection to an industrial protocol is a potential bridge between previously isolated systems.

Mitigation approaches:

  • Deploy MCP servers on edge devices within the OT DMZ, not on cloud-connected hosts
  • Use read-only MCP servers for monitoring; require separate, heavily audited write-access servers
  • Implement network segmentation between MCP servers and critical control systems
  • Never expose industrial MCP servers directly to the internet

Safety-Critical Considerations

  • Human-in-the-loop for actuator control: AI agents should never autonomously write to PLCs or execute robot programs in production without human approval. Read access for monitoring is far safer than write access for control.
  • Rate limiting: Industrial protocols have real-time constraints. An MCP server flooding a PLC with queries could disrupt control loop timing.
  • Fail-safe defaults: If an MCP connection drops, the manufacturing process must continue safely. MCP servers should be designed as advisory systems, not control dependencies.
  • Change management: Any MCP server deployment in a manufacturing environment should go through the same change management process as any other OT modification — including risk assessment, testing in a staging environment, and documented rollback procedures.

Intellectual Property Protection

Manufacturing CAD files, process parameters, and production recipes represent significant IP. MCP servers that expose this data to AI agents must consider:

  • Data exfiltration risks — AI agents with access to CAD models could inadvertently expose designs
  • API key management — industrial MCP servers need robust authentication, not just static API keys
  • Audit logging — every access to manufacturing data through MCP should be logged and reviewable
  • Role-based access — different users need different levels of access (an operator shouldn’t have design data access through MCP)

Regulatory Compliance

Manufacturing spans multiple regulatory frameworks:

  • FDA 21 CFR Part 11 for pharmaceutical and medical device manufacturing — requires electronic records and signatures
  • ISO 27001/IEC 62443 for industrial cybersecurity
  • ITAR/EAR for defense manufacturing — restricts data access by nationality
  • CE marking/UKCA for European manufacturing — traceability requirements
  • ISA-95 for manufacturing integration standards — defines boundaries between enterprise and control systems

AI agents operating across these regulatory boundaries through MCP must respect the same compliance requirements as human operators.

Ecosystem Gaps: What’s Missing

Despite 115+ MCP servers across the manufacturing ecosystem, significant gaps remain.

Major Vendor Gaps

Vendor Products MCP Status
Rockwell Automation Allen-Bradley PLCs, FactoryTalk, Plex MES No official MCP servers. EtherNet/IP covered only by NordicAgents suite
ABB ABB robots, ABB Ability, B&R Automation Only community B&R PVI server; no ABB robotics or Ability MCP
Schneider Electric Modicon PLCs, EcoStruxure, AVEVA Zero MCP presence
Honeywell Experion DCS, Forge, Safety Systems Zero MCP presence
Emerson DeltaV DCS, Fisher valves, AMS Zero MCP presence
FANUC CNC, industrial robots Zero MCP presence
KUKA Industrial robots Zero MCP presence (name overlap with kukapay/opcua-mcp is unrelated)

Technology Gaps

  • CNC machining: No MCP servers for CNC controllers (Fanuc, Siemens SINUMERIK, Haas, Mazak) — the entire CNC category is absent
  • MES platforms: No MCP servers from Siemens OpCenter, Rockwell Plex, Dassault DELMIA, or other major MES vendors (Critical Manufacturing is the sole exception, in preview)
  • PLM systems: No MCP servers for Siemens Teamcenter, PTC Windchill/Creo, or Dassault ENOVIA/3DEXPERIENCE despite these being the backbone of manufacturing product data
  • SPC/Statistical Process Control: No dedicated SPC MCP servers — quality data analysis is nearly absent
  • Manufacturing-specific ERP: No Epicor or Infor MCP servers — the ERPs most focused on manufacturing
  • Cloud IIoT platforms: No dedicated AWS IoT SiteWise or Azure IoT Hub MCP servers
  • ISA-95/ISA-88 standards: No standards-aware MCP servers that understand manufacturing data models
  • Predictive maintenance: Only early-stage implementations despite massive market demand ($17B in 2026)
  • Energy management: No manufacturing-specific energy monitoring MCP servers (ISO 50001)
  • Environmental/EHS: No EHS (Environment, Health, Safety) management MCP servers

The Siemens Exception

Siemens stands alone among industrial automation majors in embracing MCP. They have released an official MCP SDK, a developer catalog, design language integration, and official SCADA documentation — plus community servers for TIA Portal, WinCC, S7 PLCs, and Solid Edge. This first-mover advantage in MCP could become a competitive differentiator as AI-assisted manufacturing operations grow.

Getting Started by Role

Manufacturing Engineer

Start with OPC-UA or MQTT MCP servers to give AI agents read access to production data. Pair with Grafana MCP for visualization. This combination lets you ask AI questions about production trends, alarm patterns, and equipment behavior without writing dashboard queries manually.

Controls Engineer / PLC Programmer

The CODESYS MCP toolkit (37★) or TIA Portal MCP (28★) can assist with PLC programming — reviewing code, generating structured text, and navigating project structures. Start with read-only access and use AI as a code review assistant before enabling any write capabilities.

Mechanical / Design Engineer

Choose the CAD MCP server matching your platform — SolidWorks (71★), Fusion 360 (64★), FreeCAD (68★), or KiCAD for PCB work (534★). AI agents can assist with parametric design exploration, design review, and documentation generation.

Plant Manager / Operations

Grafana MCP (2,667★) combined with InfluxDB MCP (33★) gives AI agents access to your monitoring stack. Ask questions about OEE, downtime patterns, and production performance in natural language instead of building dashboard queries.

IT/OT Security Professional

Start by auditing any existing MCP servers in your environment. Review the NordicAgents IndustriConnect suite to understand what protocols AI agents could potentially access. Establish policies for MCP deployment in OT networks before engineers deploy them ad-hoc.

Supply Chain / Planning

The Dynamics 365 MCP Server (official Microsoft) or Odoo MCP covers ERP access. Combine with SupplyMaven MCP for real-time supply chain intelligence. The optim-engine MCP provides job-shop scheduling and routing optimization directly through AI conversation.

Market Context

The convergence of AI and manufacturing is accelerating across multiple dimensions:

  • AI in manufacturing: $34.18 billion (2025) → $155 billion by 2030 (35.3% CAGR)
  • Generative AI in smart manufacturing: $468 million (2026) → $5 billion by 2034 (34.5% CAGR)
  • Manufacturing execution systems: $18.6 billion (2026) → $56.7 billion by 2034 (14.9% CAGR)
  • Predictive maintenance: $17.1 billion (2026) → $97.4 billion by 2034 (24.3% CAGR)
  • Digital twin market: $49.5 billion (2026) → $328.5 billion by 2033 (16% CAGR)
  • Industrial IoT: Projected to surpass $602 billion in 2026

Adoption statistics tell a story of rapid transformation:

  • 47% global smart manufacturing adoption in early 2026 (12% year-over-year increase)
  • 40%+ of manufacturers adopting AI for scheduling systems
  • 75% of companies have adopted digital-twin technologies of medium or higher complexity
  • 70% of industrial enterprises investing in automation technology
  • AI predictive maintenance reduces unplanned downtime by 43%
  • Digital twin virtual commissioning reduces on-site commissioning time by 52%

The MCP ecosystem itself has reached significant scale: 97 million+ monthly SDK downloads, 5,800+ MCP servers, and 300+ MCP clients as of late 2025.

Key Takeaways

  1. 115+ MCP servers touch manufacturing, spanning industrial protocols, PLCs, SCADA, CAD, ERP, robotics, 3D printing, and digital twins — but coverage is uneven across categories.

  2. Siemens is the clear first-mover among industrial automation vendors, with official MCP infrastructure, an SDK, a developer catalog, and the strongest community ecosystem. Microsoft (Dynamics 365), Tulip, and Beckhoff also have official MCP investments.

  3. Industrial protocols are well-covered — OPC-UA (6 servers), MQTT (7+), and multi-protocol suites give AI agents connectivity to most factory data. NordicAgents’ IndustriConnect covers 10 protocols in one package.

  4. CAD tools have surprisingly strong coverage — SolidWorks, Fusion 360, FreeCAD, KiCAD, Onshape, AutoCAD, and Rhino all have MCP servers. KiCAD leads with 534 stars. AI-assisted design is clearly a high-demand area.

  5. Robotics integration is mature — ROS MCP servers (1,121★ for the leading implementation) enable AI-robot interaction across ROS1 and ROS2. Universal Robots cobots also have dedicated MCP servers.

  6. The biggest gaps are in the middle — MES, PLM, SPC, and CNC machining remain largely unconnected. These are the systems that contain the most manufacturing-specific knowledge and process data.

  7. Safety must come first — Manufacturing MCP deployments require different security thinking than IT applications. Start read-only, implement network segmentation, maintain human-in-the-loop for any actuator control, and treat MCP servers as OT assets subject to change management.

  8. Open-source ERP (Odoo) leads ERP coverage — with 5 MCP implementations covering manufacturing modules. SAP has community servers but no official MCP support yet.

This guide was last updated on March 29, 2026. The MCP manufacturing ecosystem is evolving rapidly — new servers appear regularly. For the latest MCP servers across all categories, see our MCP server directory. For fundamentals, start with What is MCP?.