Use MPDV HYDRA with AI via MCP
HYDRA records what actually happens on the shop floor – from data collection at the workstation to quality inspection. No vendor MCP server exists for it so far. Instead, we connect the system to CompanyGPT individually through its integration platform, together with your IT team.
What is MPDV HYDRA?
HYDRA is the manufacturing execution system built by MPDV Mikrolab in Mosbach, Germany, and one of the most widely installed MES products in the German market. It covers shop floor and machine data collection, detailed scheduling, workforce time management and quality management. The current generation, HYDRA X, is built on the Manufacturing Integration Platform (MIP) – an open platform with a semantic data model on which the individual manufacturing apps operate.
That platform architecture is what makes HYDRA interesting for AI scenarios. Production metrics are held in structured form: run times, downtime reasons, scrap rates, order confirmations. For questions from production management, maintenance or work scheduling this is a solid factual basis – but until now one you had to unlock through reports and analyses rather than a question in plain language.
Connection at a Glance
- MCP server
- None available – custom integration
- Interfaces
- Manufacturing Integration Platform (MIP), REST services
- Authentication
- Token-based via the MIP, dedicated service account
- Operating model
- On-premises, private cloud or SaaS option
- Integration hosting
- MCP server as a container in your CompanyGPT environment
- Alternative
- Connection via automation gateway (e.g. n8n)
How to Connect MPDV HYDRA to CompanyGPT
There is no MCP server for HYDRA – neither from MPDV nor from the community. The connection to CompanyGPT is therefore built as a custom integration on the existing interfaces: the Manufacturing Integration Platform has a service-based architecture and exposes REST services on top of the semantic data model. That is the natural entry point, because data from the various HYDRA modules is already consolidated there in business terms.
During the project we agree with your MES owners which objects and metrics the assistant may read, and set up a dedicated service account with token-based access for it. From the available API descriptions we build an MCP server and run it as a container inside your CompanyGPT environment. Credentials therefore stay in your tenant rather than with an external service.
HYDRA usually runs inside the production network, often deliberately segmented. We set up the link between CompanyGPT and the MES together with your IT team and your MPDV partner, aligned with your network and approval policies. Where a direct API connection is not desired, an automation gateway such as n8n can serve as an intermediate layer.
Note for on-premises operation: CompanyGPT runs in your own Azure tenant – so your on-premises network can be connected directly: via a site-to-site VPN (Azure VPN Gateway) or ExpressRoute between your corporate network and the virtual network of your CompanyGPT environment. The system does not need to be publicly reachable; we set up the connection together with your IT team.
Typical Use Cases
How AI agents in CompanyGPT work with MPDV HYDRA
Prepare the shift handover
The shift supervisor has the metrics of the finished shift summarized – output, downtime, conspicuous workstations – and uses the text as the basis for the handover.
Investigate downtime reasons
Maintenance asks which downtime reasons dominated on a machine group last month and receives the analysis with references to the underlying records.
Answer questions about an order
Sales and work scheduling check the confirmation status of an order directly in chat instead of calling the shop floor or requesting a report.
Frequently Asked Questions
MPDV HYDRA and AI – the most important answers
Does MPDV offer an MCP server for HYDRA?
No. As of August 2026 there is neither a vendor nor a community MCP server for HYDRA. Connecting AI assistants runs through the documented interfaces of the Manufacturing Integration Platform.
What is the MIP and why does it matter for the integration?
The Manufacturing Integration Platform is the open platform layer underneath HYDRA X. It provides a semantic data model and REST services that the manufacturing apps build on. For an AI connection it is the appropriate access point, because the data is already consolidated in business terms.
How does CompanyGPT reach a system inside the production network?
Through a secured network link that we set up together with your IT team. The MES does not need to be reachable from the internet, and access can be limited to the required services.
Can the assistant also write data back into HYDRA?
Technically that depends on the services you release and the rights of the service account. In practice we usually start read-only; write functions are enabled deliberately and one at a time.
Who implements the integration?
innFactory builds and operates the MCP server inside your CompanyGPT environment. Network connectivity, service accounts and the functional scope are agreed with your IT team and your MPDV contact.
MPDV HYDRA and Your AI – GDPR-Compliant with CompanyGPT
CompanyGPT is the enterprise AI assistant in your own cloud tenant: all relevant AI models, your company data and integrations like MPDV HYDRA – centrally managed, with roles and permissions. We set up the connection together with you.
MPDV HYDRA is a trademark of its respective owner (MPDV Mikrolab GmbH). Its mention describes compatibility and does not imply any partnership or endorsement. Availability and scope of the connection depend on the respective vendor.