manubes MCP Server

Analyze Production Data with Claude, ChatGPT and Copilot

The integrated MCP server makes production data from your manubes tenant available to AI assistants. Ask questions about downtime, scrap or fault reports and get analyses and recommendations based on the latest data.

  • Enabled with one click, ready to use immediately
  • Access strictly within existing manubes permissions
  • Answers from validated production data, no exports
manubes tenant setting “AI Agents, AI Chats, and Scripts (MCP)”: Access is unlocked and the MCP address is displayed
An AI assistant answers the question about the most common downtime reasons in October with a bar chart based on manubes data: Setup and material shortage cause roughly three quarters of the downtime

Integrated MCP Server

Connect manubes to AI environments

The manubes MCP Server connects AI assistants to your manubes tenant. Using the Model Context Protocol (MCP), it gives Claude, ChatGPT, Copilot and others secure, controlled access to data in manubes.

Enabled with one click

The MCP server is available for every manubes tenant. Once it is enabled in the settings, you can copy the MCP address and add it to the AI environment of your choice.

Secure sign-in via OAuth

When connecting, users confirm access with their manubes account. The AI then sees exactly what that user is allowed to see in manubes.

Your data models, readable for AI

The MCP server makes the data models and data of your tenant accessible to AI assistants. The AI knows entities, fields and relations and can query and analyze the right records in a targeted way.

How It Works

How does the manubes MCP Server work?

Diagram: Claude, ChatGPT and Microsoft Copilot access data models, process data and operator-validated data in the manubes tenant via the Model Context Protocol and an OAuth sign-in with the manubes account

You work in the AI assistant you already use, such as Claude, ChatGPT or Microsoft Copilot. Add your tenant’s MCP Server URL and authenticate the connection.

Then simply ask your questions in the chat. Whenever a question requires production data, the AI checks the data model via the MCP server and retrieves the relevant records. Answers appear directly in the chat as lists, tables or analyses. You can use the tools of your AI environment, for example to create charts or files, and combine data from manubes with information from other sources such as manuals and work instructions.

Diagram: Claude, ChatGPT and Microsoft Copilot access data models, process data and operator-validated data in the manubes tenant via the Model Context Protocol and an OAuth sign-in with the manubes account

Example Questions

What can production managers ask?

Questions are asked in the chat of the AI assistant, in plain language. The AI gathers the matching data from manubes, calculates metrics such as the reject rate and compares periods, lines, products or shifts. It summarizes long lists of entries in a few sentences. Four examples from everyday production:

Downtime

“Which line had the most downtime last month, and what were the main causes?”

The AI adds up the downtime per line and sorts it by the reasons that operators recorded on the line.
Quality

“How has the scrap rate developed over the last two quarters, broken down by product?”

The AI calculates the rate per product and quarter and shows the trend as a table or chart.
Faults

“Summarize all fault reports logged for machine 4 this week.”

The AI reads the reports including comments and summarizes recurring causes in a few sentences.
Processes

“Which checklist items are most often marked as not OK, and on which shift?”

The AI analyzes the inspection results of digital checklists and compares them between shifts.

Why It Matters

AI is only as good as the data behind it

manubes records production data cleanly and in context. Much of it is checked and enriched by operators on the line, such as the reason for a downtime event reported by the machine.

This validated process data is a reliable foundation for professional AI: Answers and analyses are based on what really happened in production.

manubes application “Downtime”: For a 49-minute downtime event detected by the machine, the operator adds the reason “Material shortage” and a comment

What are the benefits of the manubes MCP Server?

Open standard

MCP is supported by the leading AI providers. You are not tied to a single tool.

The AI you already use

Claude, ChatGPT or Microsoft Copilot. You do not need a new tool.

Current data instead of copies

The AI works directly with the data in manubes, not with outdated Excel exports.

Answers with context

The AI knows the data model and understands what the data means.

No data science project

Production managers ask their own questions and find patterns, outliers and losses themselves.

Future-proof

New AI models and tools use the same connection as soon as they appear.

Comparison

MCP server or data export:
What is the difference?

Without an MCP server, production data first has to be exported, prepared and uploaded to the AI assistant. With the manubes MCP Server, the AI accesses the data directly.

MCP server or data export: What is the difference?
Criterion Export and own analysis manubes MCP Server
Up to date As of the last export Current data from manubes
Effort Export, preparation and upload for each question Connect once, then ask
Meaning of the data Column names without context AI knows entities, fields and relations
Permissions File is passed on without access rights Existing manubes permissions apply
Who asks the question? Usually IT or data science Production managers themselves

Security

Is the manubes MCP Server secure?

The manubes MCP Server uses OAuth for authentication and authorization. Every connection runs through an existing manubes user account. Its roles and permissions apply unchanged: The AI only sees what the signed-in user is allowed to see in manubes.

MCP access can be switched off per tenant. An administrator explicitly enables it in the tenant settings. Currently, the MCP server provides read access to data.

manubes

Set up the MCP server in 4 steps

  1. Enable

    In manubes, go to Administration › Tenant Settings › “AI Agents, AI Chats, and Scripts” and enable access.

  2. Copy the address

    Copy the MCP address displayed for the tenant.

  3. Add it to your AI assistant

    Add the address in your AI assistant as a custom connector or MCP server, for example in Claude or ChatGPT.

  4. Connect and ask

    Confirm access with your manubes account and ask questions about your production in the AI assistant.

Factsheet

Download the manubes MCP Server factsheet

All features, the setup and use cases on three pages, ready to share with colleagues in IT and production.

  • How it works and setup in three steps
  • Example questions on quality, faults and processes
  • Security and permissions at a glance

Download now for free

Loading form …

Form not loading? Email cc@inray.de.

Frequently asked questions about the manubes MCP Server

What is an MCP server?

An MCP server is an interface based on the open Model Context Protocol (MCP) standard, through which AI assistants access data and tools of external systems. The manubes MCP Server provides the data models and data of a manubes tenant for this purpose.

Are MCP servers secure?

The security of an MCP server depends on authentication and permissions. The manubes MCP Server uses OAuth with existing manubes user accounts, so the AI only sees what the signed-in user is allowed to see.

Which AI assistants work with the manubes MCP Server?

The manubes MCP Server works with Claude, ChatGPT, Microsoft Copilot and other MCP-capable clients. Whether MCP support is available depends on the respective AI client and plan.

What data can the AI query via the MCP server?

Via the manubes MCP Server, the AI can read the data models and records that the signed-in user has access to in manubes. These include, for example, downtime, fault reports, inspection results and order data.

Does the manubes MCP Server require an IT project?

No. An administrator enables access in the tenant settings with one click. The MCP address is then added to the AI assistant and confirmed with the manubes account.