SAS Viya MCP Server
Official SAS Viya MCP server: run SAS code, train AutoML projects, score models and drive analytics with 115 tools
- 68 ★ on the repository
- Last commit
- Apache-2.0
Broad access
Why this level
- Runs arbitrary SAS code in the company compute environment
- Exposes corporate data and models under the user account
Install
- Works in
Manual install
claude mcp add --transport http sas-viya http://localhost:8134/mcpStart the server with `uv run app` first.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Runs arbitrary SAS code in the company compute environment
- Exposes corporate data and models under the user account
This is third-party code. Review the repository files before installing.
What it does
The server connects an agent to a SAS Viya environment. The agent runs SAS code through the compute service, reads logs and tables, trains AutoML projects and scores models. Tools are split into 11 tiers that you enable as needed, so the agent is not flooded with a long tool list. Clients with MCP Apps support (Claude, ChatGPT, Microsoft 365 Copilot, VS Code, Cursor) show interactive panels for data, SAS logs and glossary editing. Prompt templates help improve SAS code. Sign-in uses OAuth2 with PKCE; the server runs over HTTP, stdio or in a container.
Who it is for. Analysts and data scientists who already run SAS Viya and want agent access to it.
Good fit when
- You want the agent to run and debug SAS code
- You need to train an AutoML model and score it without clicking through the UI
- Your team runs SAS Viya on Kubernetes and wants a standard way to expose it to agents
- You want an interactive panel over SAS tables and logs inside Claude or VS Code
- Your team wants shared agent access to Viya through Kubernetes and an ingress
Not a fit when
- You have no SAS Viya environment, the server is useless without one
- You work with data outside SAS, such as local CSV files
- Your security policy forbids running code with a token from a local cache
Example request
Run a SAS program in Viya that computes average order value by region and show me the logMCP
- Transport
- stdio, http
- Authentication
- OAuth
Environment variables
VIYA_ENDPOINTrequired- Address of your SAS Viya environment
COMPUTE_SESSION_ID- Fixed session ID when the environment does not expose /compute/contexts
Limitations
Requires Python 3.12 or newer, uv and a SAS Viya environment with the compute service; tool availability depends on the Viya release (see COMPATIBILITY.md). Some environment setup is done by a Viya administrator. The stdio mode keeps an access token in a local cache and needs a fresh sign-in after it expires. Passing a raw bearer token is enabled by ALLOW_RAW_BEARER and VIYA_AUTH=false disables sign-in checks, so keep those flags off on shared machines.
How to disable. Remove the server entry from your client configuration and stop the process or container; the token cache lives in `~/.sas-mcp-server/credentials.json`.
FAQ
How do I run the server without Python?
Use Docker: image `ghcr.io/sassoftware/sas-mcp-server:latest`, pass the environment address in VIYA_ENDPOINT, port 8134.
How do I sign in with stdio?
Run `sas-viya auth loginCode` or `uv run sas-mcp-login` once: the token is cached locally and the client starts the server itself.
What are the 11 tool tiers?
Tools are grouped by stages of the analytics life cycle, and you enable only the groups you need to keep the agent tool list short.
Does it work with every Viya release?
No, available tools depend on the release, from LTS 2024.09 to Stable 2026.10; the table is in COMPATIBILITY.md.
Where is the token stored?
In stdio mode it goes to `~/.sas/credentials.json` or `~/.sas-mcp-server/credentials.json`; no password is stored.
Discussion
No questions or reviews yet.
Related
MCP serversvi-sql is a terminal SQL client with a built-in MCP server: the agent sees the schema and drafts queries straight in your editor
An MCP server for Microsoft SQL Server and Azure SQL: the agent lists tables and runs SQL queries that read and write data
MCP server and CLI to drive real Microsoft Excel: Power Query, DAX, VBA, PivotTables and charts
One MCP server for SQLite, SQL Server, PostgreSQL and MySQL: read and change data, edit schemas and export to CSV and JSON
MCP server for SAP HANA and HANA Cloud: schema browsing, SQL with row caps and 34 tools, with writes blocked by default
Official Tableau MCP server: the agent finds data sources, sends queries to them and opens dashboards and Pulse metrics
README in short
The repository covers installation with Python and uv, running in HTTP, stdio, Docker and Kubernetes modes, OAuth2 setup and environment variables. SAS articles on connecting GitHub Copilot and Claude Code are linked.