AI & Analytics Legends The knowledge platform for SAP Analytics Sign inSign up
Guide

SAP Analytics Cloud MCP: servers, scope and production checks

SAP Analytics Cloud MCP: servers, scope and production checks — Analytics Legends section illustration for SAP analytics market intelligence (Datasphere, BDC, SAC, BW/4HANA)

As of 2026-10-18

As of 18 October 2026 our verified catalogue lists no official SAP MCP server for SAP Analytics Cloud. What exists is one community server, published under an MIT licence, that wraps the SAC REST API as 90 tools. It is a usable starting point for experiments, not something to put in front of finance users without a security review.

What exists, and who maintains it

The only SAC-specific MCP server in our catalogue is the community project "SAP Analytics Cloud MCP" (GitHub, MIT licence). It is not an SAP product and carries no SAP support commitment. Its status label and transport are not stated in the catalogue entry, so check the repository before assuming either stdio or HTTP.

We also searched the SAP Help Portal for an official SAC, Datasphere or HANA Cloud MCP server and found no such page. That is a statement about 18 October 2026 only; SAP has been adding MCP servers quickly in other areas, so re-check before you rule it out for a project.

What it lets an assistant do

The server wraps the SAC REST API as 90 tools, authenticated with OAuth client credentials. It adds a smart-query tool and guided prompts on top. In practice that means an assistant can discover and call SAC administration and content endpoints through tool calls instead of you writing scripts.

The catalogue does not say which of the 90 tools read and which write. Treat that as unknown until you have listed the tools yourself. A tool surface of that size is also a context cost: many clients degrade when a server advertises dozens of tools, so plan to restrict the exposed set.

Wiring it up

Create an OAuth client in SAC for the integration, store its credentials outside the repository, and run the server where your MCP client can reach it. Client credentials describe the integration, not the person asking the question, so every call runs under the same technical identity whoever is chatting.

That is the main design point for a SAC estate. Story and model access in SAC is driven by roles and teams; a single technical client can see far more than any individual analyst should. Give that client the smallest role that covers your use case, and keep it out of tenants holding planning data until you have decided who may query it.

What to check before production

Identity: client credentials flatten user-level security, so decide whether answers are allowed to bypass row-level restrictions you set in SAC. Scope: list every tool and disable anything that creates, changes or deletes content. Audit: confirm that the MCP server logs the tool call and that SAC logs the API call, so you can tie a change back to a conversation. Supply chain: a 90-tool community project needs the same review as any third-party dependency, with a named owner and an exit plan.

The SAP-side rule also matters. SAP's API policy, as summarised in our concept notes, limits agentic use of SAP APIs to SAP-endorsed architectures and warns against bypassing them through proxies or gateways. Ask your SAP account team whether a community wrapper around the SAC REST API sits inside that perimeter for your contract.

The realistic alternatives

If the goal is a governed route, the endorsed pattern is SAP Integration Suite: its MCP gateway puts SAP and non-SAP APIs behind one entry point with authentication, rate limiting and monitoring, and it needs the Premium or Enhanced edition plus the Integration Cell runtime. The catalogue does not describe a ready-made SAC artefact there, so you would model the SAC APIs you need yourself.

For end users, SAC already has its own assistant: Joule in SAP Analytics Cloud answers questions inside the product, under the user's own authorisations. That is not MCP, and it is often the better answer for business users. MCP earns its place when you want an external assistant such as a coding agent or a custom agent to reach SAC alongside other systems.

For market questions rather than tenant data, such as which firms deliver SAC projects or what SAC contract rates look like, the Analytics Legends MCP serves read-only public tools; it does not connect to your SAC tenant.

What we cannot assert

Transport, status and the read/write split of the community server's 90 tools are not stated in the catalogue; a future official SAP server may supersede this page.

Frequently asked

Is there an official SAP MCP server for SAP Analytics Cloud?

Not in our verified catalogue of 18 October 2026. The SAP Help Portal search returned no official SAC, Datasphere or HANA Cloud MCP page. The only SAC-specific server found is a community one.

Can an assistant change my SAC stories through MCP?

The catalogue does not say which of the community server's 90 tools write. Assume it can until you have inspected the tool list, and restrict the OAuth client's role accordingly.

Does Joule in SAC use MCP?

Joule in SAC is the product's built-in assistant and works under the user's own authorisations. Our sources do not describe it as an MCP server. MCP is a separate route for external assistants.

Should I use the community server in production?

Only after review: owner, licence, tool scope, credentials handling, logging and an exit plan. For governed access, model the SAC APIs behind the Integration Suite MCP gateway instead.

Engage the editor

This guide is written by . 27 years of SAP — Datasphere, Business Data Cloud, SAC Planning, Joule. Freelance or permanent · Hybrid / remote · Anywhere in EMEA. Available from 19 October 2026.

Profile and contact

What this page is built on

External sources

Read next