MSSQL MCP Server
Provides read-only access to Microsoft SQL Server databases using Windows Authentication, enabling AI assistants to safely explore schemas and query data with built-in security controls.
README
pyodbc MCP Server
A Model Context Protocol (MCP) server that provides read-only access to Microsoft SQL Server databases using Windows Authentication.
Built for environments where:
- 🔐 Windows Authentication is required (no username/password storage)
- 🛡️ Read-only access is mandated by IT security policy
- 🖥️ SQL Server is accessed from Windows workstations
- 🤖 AI assistants need safe database access (Claude Code, etc.)
Features
- Windows Authentication - Uses
Trusted_Connectionvia pyodbc, no credentials to manage - Read-only by design - Only SELECT queries allowed, dangerous keywords blocked
- Row limiting - Prevents accidental large result sets (configurable, max 1000)
- Schema exploration - List tables, views, describe columns, find relationships
- MCP compatible - Works with Claude Code, Claude Desktop, and any MCP client
Available Tools
| Tool | Description |
|---|---|
ListTables |
List all tables in the database, optionally filtered by schema |
ListViews |
List all views in the database, optionally filtered by schema |
DescribeTable |
Get column definitions for a specific table |
GetTableRelationships |
Find foreign key relationships for a table |
ReadData |
Execute a SELECT query (with security filtering) |
Installation
Prerequisites
- Python 3.10+
- Windows with ODBC Driver 17+ for SQL Server
- Network access to your SQL Server
- Windows domain account with SELECT permissions on target database
Install from PyPI
pip install pyodbc-mcp-server
Install from Source
git clone https://github.com/jjones-wps/pyodbc-mcp-server.git
cd pyodbc-mcp-server
pip install -e .
Install ODBC Driver (if needed)
Download and install Microsoft ODBC Driver 17 for SQL Server.
Configuration
Environment Variables
| Variable | Default | Description |
|---|---|---|
MSSQL_SERVER |
localhost |
SQL Server hostname or IP |
MSSQL_DATABASE |
master |
Target database name |
ODBC_DRIVER |
ODBC Driver 17 for SQL Server |
ODBC driver name |
Claude Code Configuration
Add to your ~/.claude.json (or %USERPROFILE%\.claude.json on Windows):
{
"mcpServers": {
"mssql": {
"type": "stdio",
"command": "cmd",
"args": ["/c", "python", "-m", "mssql_mcp_server"],
"env": {
"MSSQL_SERVER": "your-sql-server",
"MSSQL_DATABASE": "your-database"
}
}
}
}
Alternative: Direct script execution
{
"mcpServers": {
"mssql": {
"type": "stdio",
"command": "cmd",
"args": [
"/c",
"python",
"C:\\path\\to\\pyodbc-mcp-server\\src\\mssql_mcp_server\\server.py"
],
"env": {
"MSSQL_SERVER": "your-sql-server",
"MSSQL_DATABASE": "your-database"
}
}
}
}
Note for Windows users: The
cmd /cwrapper is required for proper stdio communication with MCP clients on Windows.
Claude Desktop Configuration
Add to your Claude Desktop config (%APPDATA%\Claude\claude_desktop_config.json):
{
"mcpServers": {
"mssql": {
"command": "python",
"args": ["-m", "mssql_mcp_server"],
"env": {
"MSSQL_SERVER": "your-sql-server",
"MSSQL_DATABASE": "your-database"
}
}
}
}
Usage Examples
Once configured, you can ask Claude to:
Explore Schema
"List all tables in the dbo schema"
"Describe the structure of the customers table"
"What are the foreign key relationships for the orders table?"
Query Data
"Show me the first 10 rows from the products table"
"Find all orders from the last 30 days"
"What are the top 5 customers by total order value?"
Analyze Relationships
"Find all tables that reference the customer table"
"Show me the relationship between orders and order_lines"
Security
This server is designed with security as a primary concern:
Read-Only Enforcement
- Only queries starting with
SELECTare allowed - Dangerous keywords are blocked even in subqueries:
INSERT,UPDATE,DELETE,DROP,CREATE,ALTEREXEC,EXECUTE,TRUNCATE,GRANT,REVOKE,DENYBACKUP,RESTORE,SHUTDOWN,DBCC
Windows Authentication
- Uses
Trusted_Connection=yes- no passwords stored or transmitted - Leverages existing Windows domain security
- Your database permissions are enforced by SQL Server
Row Limiting
- Default limit: 100 rows per query
- Maximum limit: 1000 rows per query
- Prevents accidental retrieval of large datasets
Development
Running Tests
pip install -e ".[dev]"
pytest
Running Locally
# Set environment variables
export MSSQL_SERVER=your-server
export MSSQL_DATABASE=your-database
# Run the server
python -m mssql_mcp_server
Troubleshooting
"ODBC Driver not found"
Install the Microsoft ODBC Driver for SQL Server:
"Login failed" or "Cannot connect"
- Verify your Windows account has access to the SQL Server
- Test connection with
sqlcmd -S your-server -d your-database -E - Check firewall allows connection on port 1433
"Tools not appearing in Claude Code"
- Ensure
type: "stdio"is in your config - Use the
cmd /cwrapper on Windows - Restart Claude Code after config changes
- Check Claude Code logs for MCP errors
Contributing
Contributions are welcome! Please:
- Fork the repository
- Create a feature branch
- Add tests for new functionality
- Submit a pull request
License
MIT License - see LICENSE file.
Acknowledgments
- Built with FastMCP for MCP protocol handling
- Uses pyodbc for SQL Server connectivity
- Inspired by the need for safe AI access to enterprise databases
Related Projects
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.