karthickshiva-png/sap-cpi-mcp-server

MCP server for SAP Cloud Integration - monitor Message Processing Logs and manage integration runtime artifacts using natural language

About sap-cpi-mcp-server

An MCP (Model Context Protocol) implementation for SAP Cloud Integration (CPI) using SAP Integration Suite's native MCP Server artifact type and OpenAPI.

The project was built and tested with VS Code Copilot as the MCP client.

The current design intentionally keeps Monitoring and Runtime Operations on the standard CPI OData API, while Deploy is handled by a small wrapper iFlow.

The MCP server exposes Message Processing Log (MPL) operations such as:

- List message processing logs - Filter logs by status/time using OData query options - Sort results - Fetch a specific log by Message GUID - Get message counts - Expand adapter attributes/custom header properties where supported by the API

The MCP server exposes runtime artifact operations such as:

- List deployed integration artifacts - Check an artifact's runtime status - Get an artifact by ID - Undeploy an artifact

Deployment is deliberately implemented through a separate wrapper iFlow.

The wrapper accepts a clean JSON request:

and internally calls SAP Cloud Integration's OData v2 DeployIntegrationDesigntimeArtifact FunctionImport.

The standard CPI deployment API is an OData v2 FunctionImport. When it was exposed directly as an MCP tool, the MCP client could invoke the tool, but the deployment request failed with an error similar to:

The issue was encountered while mapping the FunctionImport's query-string parameters through the MCP tool invocation.

From the project README.

Maintaining this server?

Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:

[![On MCP Radar](https://mcp.liqiwa.com/badge.svg)](https://mcp.liqiwa.com/s/karthickshiva-png--sap-cpi-mcp-server.html)