Micro Analytic Service (MAS) - Model Context Protocol (MCP) Tool Maker

A Model Context Protocol (MCP) Server Script for your SAS Viya Micro Analytic Service (MAS) modules

This repository contains a Python package that creates a Model Context Protocol (MCP) server script for calling published SAS Micro Analytic Services (MAS) modules discovered in a SAS Viya environment. The output is a Python script (in two flavours) for an MCP server served through stdio, Streamable HTTP (planned) and a format customised for SAS Retrieval Agent Manager (RAM).

Python package at: https://pypi.org/project/mas-mcp-toolmaker/

A wiki of this repo has been generated using DeepWiki and is available here:

Note that this open-source package is provided as a convenient way to interface with a SAS Viya environment over MCP. SAS Viya is proprietary software from SAS Institute available in several offerings. Running SAS Viya requires a license and functioning credentials.

📢 Watch this video to learn more.

Installation & Set up

1. Requirements: Python 3.12 or later

2. Access to a SAS Viya environment with SAS Micro Analytic Service (MAS). This is typically a SAS Viya Enterprise or SAS Viya Advanced offering. Refer this page for details.

3. While not mandatory, we recommended standing up a virtual environment before installing this package. A convenient shell script build.sh is provided for this purpose. Refer this repository for a starter build script. Also read this article for more details.

4. The simplest way is to pip install this package from pypi.org

pip install mas-mcp-toolmaker

5. In case you would like to perform a local install,

pip install -e .

6. Use a .env file (refer sample.env, create a copy and rename to .env) to set environment variables for the following: (Refer this link for more details on obtaining access tokens: Authentication to SAS Viya: a couple of approaches)
   • VIYA_HOST: A URL pointing to your SAS Viya environment
   • VIYA_ACCESS_TOKEN: An access token which helps you authenticate to SAS Viya

Quick Start: Running the notebook

# if using virtual environment (use Windows equivalent where applicable)
cd build
. buildproj/bin/activate
jupyter-lab

The notebook mas-mcp-toolmaker.ipynb contains further details.

Quick Start: Basic Commands

Using this package is very simple. At a fundamental level, there are two operations involved.

1. Generating a list of MAS modules from a SAS Viya environment

from mas_mcp_toolmaker import MASModule

masm = MASModule()

2. Writing the extracted modules in tool representation to an MCP server script of your chosen flavour.

masm.script_stdio_server("your_server_file.py")

# or, if you wish to write to a Retrieval Agent Manager file 

masm.script_ram_server("your_ram_server_file.py")

Refer the notebook for other functions.

Contact

• Sundaresh Sankaran (email)

Version

• Version 1.2.1 (04JUL2026)
• Version 1.2.0 (03JUL2026)
• Version 1.0.0 (13MAY2026)

Refer CHANGELOG.md for a detailed change log.