Connect ChatGPT and Claude to your ERP with an MCP server.

In shortAn MCP server is a small, secure service that exposes your company’s data and actions as tools that AI assistants such as Claude and ChatGPT can call using the Model Context Protocol. Kodenova builds MCP servers for SAP Business One, SAP HANA, SQL Server and custom systems, with authentication, permissions and logging.
What’s included

What you get.

ERP tools for AI

Expose “get customer balance”, “check stock”, “create sales quotation (draft)” as safe, typed tools.

Remote (HTTP) or local MCP

Hosted MCP servers with OAuth for teams, or local servers for single-user setups.

Permissions and audit

Per-user scopes, read-only defaults and a full log of every tool call.

Works with major AI clients

Claude, ChatGPT and other MCP-compatible clients and agent frameworks.

Hosting

Deploy on your server, your cloud, or a managed Windows/Linux VPS we set up.

Documentation

Tool descriptions written so AI models call the right tool with the right inputs.

Who it’s for

Built for teams like yours.

01

Companies that want staff to use Claude or ChatGPT with live business data

02

Software vendors who want their product to work inside AI assistants

03

SAP partners building AI offers for their customers

Model Context ProtocolTypeScriptC# / .NETPythonOAuth 2.1SAP B1 Service LayerSAP HANASQL ServerCloudflareWindows Server
Process

How a project runs.

  1. Discovery call (free): we understand the goal, your SAP/ERP version and constraints.
  2. Written proposal: scope, technical approach, timeline and fixed price per phase.
  3. Build with weekly demos: you see working software early, not just status reports.
  4. Test and go-live: testing in your environment, training and documentation.
  5. 3 months of free support: fixes and small adjustments after go-live, then optional monthly support.
FAQ

MCP server development: common questions

What is the Model Context Protocol (MCP)?

MCP is an open protocol, introduced by Anthropic in 2024, that standardises how AI applications connect to external tools and data. An MCP server describes tools; an AI client like Claude or ChatGPT calls them when needed.

Why use MCP instead of a custom chatbot?

With MCP, your data becomes available inside the AI tools your team already uses, and the same server works across multiple AI clients instead of being locked into one chatbot.

Is it safe to connect AI to our ERP through MCP?

It is safe when the server enforces authentication, per-user permissions, read-only defaults and logging. Write actions should create drafts that a human approves. That is how we build them.

Have you built MCP servers before?

Yes. We built an MCP server for our SAP Business One AI connector and use MCP daily in our own development and marketing work.

Ready to get started?

Tell us what you need. You get a clear technical answer, a fixed scope and a price, usually within 24 hours.