Use proALPHA with AI via MCP
proALPHA runs production, materials and finance for manufacturing mid-market companies – usually from their own data center. An open-source MCP server exists for AI access, which we operate as a container inside your CompanyGPT environment; alternatively we build directly on the ray API.
What is proALPHA?
proALPHA is a complete ERP solution from proALPHA GmbH, based in Weilerbach, Germany, and aimed at mid-sized manufacturers. Its scope spans production planning and control, materials management and purchasing through to finance, project management and quality assurance. The system is classically deployed on-premises, while the vendor gradually expands its cloud and private cloud offerings.
Manufacturing in particular generates data whose interdependencies are hard to keep track of day to day: production orders, bills of material, stock, dates and costs are tightly coupled. An AI agent with read access can lay out those relationships on demand – for instance when work preparation or sales want to understand why a delivery date is slipping.
Connection at a Glance
- MCP server
- Community (unofficial), self-hosted in your environment
- Repository
- NinjaEde/mcp-proalpha (Python)
- Alternative interface
- proALPHA ray API (REST/OpenAPI), Integration Workbench
- Authentication
- OAuth 2.0 / token on the ray API, or a service user
- Operating model
- Mostly on-premises, increasingly private cloud
- Note
- Read and write scope defined per object beforehand
How to Connect proALPHA to CompanyGPT
proALPHA does not offer an official MCP server. The community project NinjaEde/mcp-proalpha provides an unofficial MCP server implemented in Python. Where that server is a good fit, we run it as a container inside your CompanyGPT environment – queries to your ERP stay within your own infrastructure and no externally operated service is involved.
For production scenarios the second route is often more robust: a custom integration built on the proALPHA ray API, a REST interface with an OpenAPI description, complemented by the Integration Workbench (INWB) as the established integration layer. Since we make any API with an OpenAPI specification available as an MCP integration, this yields a cleanly scoped tool set for your agents.
Authentication uses OAuth 2.0 or tokens on the ray API; classic setups work with a service user through the Integration Workbench. Which objects may be read, and whether writing is allowed at all, is defined up front. In CompanyGPT your administrators then decide by role and group which departments use the proALPHA integration. Connectivity to the system running on your network is arranged with your IT team.
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 proALPHA
Clarify production order status
Work preparation asks about status, backlog and material availability for a production order and gets the relevant figures in one answer instead of across several screens.
Assess material requirements
Purchasing compares stock, open purchase orders and planned demand for a part group to spot critical items earlier.
Summarize project status
Project management condenses effort, progress and open items of a customer project into a status note for the weekly meeting.
Frequently Asked Questions
proALPHA and AI – the most important answers
Does proALPHA have an official MCP server?
No. An unofficial community server exists (NinjaEde/mcp-proalpha), which we operate ourselves where suitable. Alternatively we build the integration directly on the proALPHA ray API.
Where does the MCP server run?
As a container inside your own CompanyGPT environment. The connection to the ERP therefore stays within your infrastructure, without an additional external provider.
What is the ray API?
proALPHA's REST interface with an OpenAPI description. It is a good basis for an MCP integration because objects and fields can be addressed and restricted precisely.
How does CompanyGPT reach our on-premises system?
Through a secured network connection between your CompanyGPT environment and the proALPHA server. We set this up with your IT team – public exposure of the ERP is not required.
Can the AI create or change data in proALPHA?
Only within what you approve. Many projects start read-only; write capabilities are added selectively for individual, clearly defined workflows.
Related Integrations
ERP & Merchandise Management
proALPHA 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 proALPHA – centrally managed, with roles and permissions. We set up the connection together with you.
proALPHA is a trademark of its respective owner (proALPHA GmbH). Its mention describes compatibility and does not imply any partnership or endorsement. Availability and scope of the connection depend on the respective vendor.