SAP / MODEL CONTEXT PROTOCOL

Controlled SAP context
for AI agents

NeKu AI SAP MCP Server exposes SAP documentation and permitted action tools to AI agents through the Model Context Protocol.

Request a demo
NEKU AI / 2026ACTIVE
MCP / SAP
  1. 01DOCS
  2. 02ACTIONS
  3. 03AUDIT

01 / SHORT ANSWER

What is an SAP MCP Server?

MCP stands for Model Context Protocol. NeKu AI SAP MCP Server creates a standard tool and context layer between AI clients and SAP systems. An agent can read exposed SAP documentation or invoke only the SAP tools it is permitted to use.

MCP does not replace SAP authorization. The target system, user identity, access mode, audit and output-redaction rules remain control layers for every connection.

Protocol
Model Context Protocol
Tool groups
SAP Docs and SAP Actions
System selection
Explicit context on every call

02 / TWO TOOL BOUNDARIES

Reading information and executing actions remain separate

SAP DOCS

Access to documentation and context

Lets agents research, explain and draft content using SAP documentation and authorized knowledge sources.

  • Knowledge access
  • Context-grounded answers
  • Documentation-focused
SAP ACTIONS

Permission-controlled SAP tools

Lets agents invoke defined query or action tools in a target SAP system within user authorization and access-mode limits.

  • Multiple SAP systems
  • Per-user identity
  • Read and action modes

03 / CONTROL LAYERS

Context and boundaries on every tool call

The MCP connection is more than a technical bridge; it carries the system, user and permission boundaries of the workflow.

01

Multi-system selection

ECC and S/4HANA environments are separated as named connections and each call identifies the target.

02

Per-user credentials

SAP connections are assigned to users and credentials are never exposed in client responses.

03

Access modes

read_only, development and full_control modes restrict the tools that can be invoked.

04

Audit and redaction

Tool calls are auditable, while sensitive fields and technical secrets can be removed from output.

04 / COMPATIBILITY

Designed around the environments SAP teams operate

ECC

SAP ECC EHP8

A supported connection target for bringing AI tools to established ECC environments.

1809

SAP S/4HANA 1809

A supported release for S/4HANA transformation and operations scenarios.

2025

SAP S/4HANA 2025

A supported connection target for current S/4HANA environments.

DEP

Cloud and on-premise NeKu AI

The MCP layer is positioned within the selected NeKu AI deployment and enterprise architecture.

05 / FREQUENTLY ASKED QUESTIONS

Short answers about SAP MCP

01What does MCP stand for?

MCP stands for Model Context Protocol, a standard way for AI clients to work with tools and context sources.

02Are SAP MCP and the n8n SAP Connector the same?

No. SAP MCP exposes context and tools to AI agents. The n8n SAP Connector links SAP data and actions to n8n workflows.

03Can SAP MCP connect to multiple SAP systems?

Yes. Connections can be managed as named systems and each tool call explicitly selects its target system.

04Are SAP credentials sent to the AI model?

No. SAP credentials stay in the connection layer and are not returned to the model or client.

05Is read-only use supported?

Yes. read_only mode disables tools that modify data and limits the connection to permitted read operations.

LET'S DESIGN YOUR PROCESSES TOGETHER

Make your enterprise workflows autonomous.

Request a demo
NEKU/SEARCH

Search the NeKu AI knowledge network