A comprehensive Microsoft SQL Server client implementing the Model Context Protocol (MCP). This server provides extensive SQL Server capabilities including query execution, schema discovery, and stored procedure management through a simple MCP interface.
The SQL Server MCP client is built with .NET Core using the Model Context Protocol C# SDK (github.com/modelcontextprotocol/csharp-sdk). It provides tools for executing SQL queries, managing stored procedures, listing tables, and retrieving comprehensive schema information from SQL Server databases. The server is designed to be lightweight yet powerful, demonstrating how to create a robust MCP server with practical database functionality. It can be deployed either directly on a machine or as a Docker container.
The MCP client operates in one of two modes:
- Database Mode: When a specific database is specified in the connection string, only operations within that database context are available
- Server Mode: When no database is specified in the connection string, server-wide operations across all databases are available
- Execute SQL queries on connected SQL Server databases
- List all tables with schema and row count information
- Retrieve detailed schema information for specific tables
- Comprehensive stored procedure management and execution
- Export/import tables as CSV
- Paginated and streaming query support
- Index, view, and function metadata
- Script execution and discovery tools
- Cursor guidance for query navigation
- Configurable tool enablement for security
- Environment-based configuration
- Comprehensive error handling and validation
- .NET 8.0 (for local development/deployment)
- Docker (for container deployment)
-
Clone this repository:
git clone https://github.com/akashCF/dotnet-mssql-mcp.git
-
Navigate to the project root directory:
cd dotnet-mssql-mcp
-
Build the project using:
dotnet build src/mssqlclient.sln
The SQL Server MCP Client is available on Docker Hub.
# Pull the latest version
docker pull akashcf/dotnet-mssql-mcp:latest
If you need to build the Docker image yourself:
# Navigate to the repository root
cd dotnet-mssql-mcp
# Build the Docker image
docker build -f src/Core.Infrastructure.McpServer/Dockerfile -t dotnet-mssql-mcp:latest src/
# Run the locally built image
docker run -d --name mssql-mcp -e "MSSQL_CONNECTIONSTRING=Server=your_server;Database=your_db;User Id=your_user;Password=your_password;TrustServerCertificate=True;" dotnet-mssql-mcp:latest
To connect to the SQL Server MCP Client from your applications:
- Use the Model Context Protocol C# SDK or any MCP-compatible client
- Configure your client to connect to the server's endpoint
- Call the available tools described below
The available tools differ depending on which mode the server is operating in, with some tools available in both modes:
Returns detailed information about the capabilities and features of the connected SQL Server instance.
Example request:
{
"name": "server_capabilities",
"parameters": {}
}
Example response in Server Mode:
{
"version": "Microsoft SQL Server 2019",
"majorVersion": 15,
"minorVersion": 0,
"buildNumber": 4123,
"edition": "Enterprise Edition",
"isAzureSqlDatabase": false,
"isAzureVmSqlServer": false,
"isOnPremisesSqlServer": true,
"toolMode": "server",
"features": {
"supportsPartitioning": true,
"supportsColumnstoreIndex": true,
"supportsJson": true,
"supportsInMemoryOLTP": true,
"supportsRowLevelSecurity": true,
"supportsDynamicDataMasking": true,
"supportsDataCompression": true,
"supportsDatabaseSnapshots": true,
"supportsQueryStore": true,
"supportsResumableIndexOperations": true,
"supportsGraphDatabase": true,
"supportsAlwaysEncrypted": true,
"supportsExactRowCount": true,
"supportsDetailedIndexMetadata": true,
"supportsTemporalTables": true
}
}
This tool is useful for:
- Determining which features are available in your SQL Server instance
- Debugging compatibility issues
- Understanding which query patterns will be used
- Verifying whether you're in server or database mode
Provides guidance for navigating large result sets using cursors.
Example request:
{
"name": "cursor_guide",
"parameters": { "query": "SELECT * FROM Orders" }
}
Executes a query and returns results in pages.
Example request:
{
"name": "paginated_query",
"parameters": { "query": "SELECT * FROM Orders", "page": 1, "pageSize": 100 }
}
Streams large query results in chunks.
Example request:
{
"name": "query_streamer",
"parameters": { "query": "SELECT * FROM Orders" }
}
Exports a table to CSV format.
Example request:
{
"name": "export_table_csv",
"parameters": { "tableName": "Customers" }
}
Imports data from a CSV file into a table.
Example request:
{
"name": "import_table_csv",
"parameters": { "tableName": "Customers", "csvData": "...csv content..." }
}
Executes a SQL script.
Example request:
{
"name": "run_script",
"parameters": { "script": "CREATE TABLE Test (Id INT)" }
}
Discovers schema, tables, and other metadata.
Example request:
{
"name": "discover",
"parameters": {}
}
Lists all tables in all databases.
Example request:
{
"name": "server_list_tables",
"parameters": { "databaseName": "Northwind" }
}
Lists all databases on the server.
Example request:
{
"name": "server_list_databases",
"parameters": {}
}
Gets the schema for a table in a specific database.
Example request:
{
"name": "server_get_table_schema",
"parameters": { "databaseName": "Northwind", "tableName": "Customers" }
}
Lists all stored procedures in a database.
Example request:
{
"name": "server_list_stored_procedures",
"parameters": { "databaseName": "Northwind" }
}
Gets the SQL definition of a stored procedure in a database.
Example request:
{
"name": "server_get_stored_procedure_definition",
"parameters": { "databaseName": "Northwind", "procedureName": "GetCustomerOrders" }
}
Gets parameter info for a stored procedure in a database.
Example request:
{
"name": "server_get_stored_procedure_parameters",
"parameters": { "databaseName": "Northwind", "procedureName": "CreateNewCustomer" }
}
Executes a query in a specific database (if enabled).
Example request:
{
"name": "server_execute_query",
"parameters": { "databaseName": "Northwind", "query": "SELECT * FROM Customers" }
}
Executes a stored procedure in a specific database (if enabled).
Example request:
{
"name": "server_execute_stored_procedure",
"parameters": { "databaseName": "Northwind", "procedureName": "CreateNewCustomer", "parameters": { "CompanyName": "Acme Corp" } }
}
Returns metadata about functions in a database.
Example request:
{
"name": "get_function_details",
"parameters": { "databaseName": "Northwind" }
}
Returns metadata about views in a database.
Example request:
{
"name": "get_view_details",
"parameters": { "databaseName": "Northwind" }
}
Returns metadata about indexes in a database.
Example request:
{
"name": "get_index_details",
"parameters": { "databaseName": "Northwind" }
}
Lists all tables in the connected database.
Example request:
{
"name": "list_tables",
"parameters": {}
}
Gets the schema for a table in the connected database.
Example request:
{
"name": "get_table_schema",
"parameters": { "tableName": "Customers" }
}
Lists all stored procedures in the connected database.
Example request:
{
"name": "list_stored_procedures",
"parameters": {}
}
Gets the SQL definition of a stored procedure in the connected database.
Example request:
{
"name": "get_stored_procedure_definition",
"parameters": { "procedureName": "GetCustomerOrders" }
}
Gets parameter info for a stored procedure in the connected database.
Example request:
{
"name": "get_stored_procedure_parameters",
"parameters": { "procedureName": "CreateNewCustomer" }
}
Executes a query in the connected database (if enabled).
Example request:
{
"name": "execute_query",
"parameters": { "query": "SELECT * FROM Customers" }
}
Executes a stored procedure in the connected database (if enabled).
Example request:
{
"name": "execute_stored_procedure",
"parameters": { "procedureName": "CreateNewCustomer", "parameters": { "CompanyName": "Acme Corp" } }
}
- src/mssqlclient.sln: Main solution file.
- Core.Infrastructure.McpServer/: Main server implementation and entry point (
Program.cs
). - Core.Application/: Core application logic and interfaces.
- Core.Infrastructure.SqlClient/: SQL Server-specific data access and services.
- UnitTests.Infrastructure.McpServer/, UnitTests.Infrastructure.SqlClient/: Unit tests for server and SQL client logic.
- scripts/: Utility scripts for versioning and maintenance.
- cleanup-tools.sh, cleanup-tools.bat: Scripts to clean up build artifacts and tools.
- scripts/create-version.ps1: PowerShell script for version management.
- cleanup-tools.sh, cleanup-tools.bat: Clean up build artifacts and temporary files.
The server provides granular control over which potentially dangerous operations are available:
By default, SQL query execution tools are disabled for security reasons. To enable these tools, set the EnableExecuteQuery
configuration setting to true
.
By default, stored procedure execution tools are disabled for security reasons. To enable these tools, set the EnableExecuteStoredProcedure
configuration setting to true
.
These can be configured in several ways:
- In the
appsettings.json
file:
{
"EnableExecuteQuery": true,
"EnableExecuteStoredProcedure": true
}
- As environment variables when running the container:
docker run \
-e "EnableExecuteQuery=true" \
-e "EnableExecuteStoredProcedure=true" \
-e "MSSQL_CONNECTIONSTRING=Server=your_server;..." \
akashcf/dotnet-mssql-mcp:latest
- In the Claude Desktop configuration:
"mssql": {
"command": "dotnet",
"args": [
"YOUR_PATH_TO_DLL\\Core.Infrastructure.McpServer.dll"
],
"env": {
"MSSQL_CONNECTIONSTRING": "Server=your_server;...",
"EnableExecuteQuery": "true",
"EnableExecuteStoredProcedure": "true"
}
}
- In the VSCode configuration:
"mssql": {
"type": "stdio",
"command": "dotnet",
"args": [
"exec",
"YOUR_PATH_TO_DLL/bin/Debug/net8.0/Core.Infrastructure.McpServer.dll"
],
"env":{
"MSSQL_CONNECTIONSTRING": "Data Source=yourserver;Initial Catalog=databasename;uid=user;Password=*****;Encrypt=False;TrustServerCertificate=False;ApplicationIntent=ReadWrite;MultiSubnetFailover=False"
}
}
When these settings are false
(the default), the respective execution tools will not be registered and will not be available to clients. This provides additional security layers when you only want to allow read-only operations.
The SQL Server connection string is required to connect to your database. This connection string should include server information, authentication details, and any required connection options.
You can set the connection string using the MSSQL_CONNECTIONSTRING
environment variable:
# Database Mode with both execution types enabled
docker run \
-e "EnableExecuteQuery=true" \
-e "EnableExecuteStoredProcedure=true" \
-e "MSSQL_CONNECTIONSTRING=Server=your_server;Database=your_db;User Id=your_user;Password=your_password;TrustServerCertificate=True;" \
akashcf/dotnet-mssql-mcp:latest
# Server Mode with both execution types enabled
docker run \
-e "EnableExecuteQuery=true" \
-e "EnableExecuteStoredProcedure=true" \
-e "MSSQL_CONNECTIONSTRING=Server=your_server;User Id=your_user;Password=your_password;TrustServerCertificate=True;" \
akashcf/dotnet-mssql-mcp:latest
The MCP server automatically detects the mode based on the connection string:
- Server Mode: When no database is specified in the connection string (no
Database=
orInitial Catalog=
parameter) - Database Mode: When a specific database is specified in the connection string
Example connection strings:
# Database Mode - Connects to specific database
Server=database.example.com;Database=Northwind;User Id=sa;Password=YourPassword;TrustServerCertificate=True;
# Server Mode - No specific database
Server=database.example.com;User Id=sa;Password=YourPassword;TrustServerCertificate=True;
# Database Mode with Windows Authentication
Server=database.example.com;Database=Northwind;Integrated Security=SSPI;TrustServerCertificate=True;
# Server Mode with specific port
Server=database.example.com,1433;User Id=sa;Password=YourPassword;TrustServerCertificate=True;
If no connection string is provided, the server will return an error message when attempting to use the tools.
Integrated security will not work from a docker container!
To configure Claude Desktop to use a locally installed SQL Server MCP client:
- Add the server configuration to the
mcpServers
section in your Claude Desktop configuration:
"mssql": {
"command": "dotnet",
"args": [
"YOUR_PATH_TO_DLL\\Core.Infrastructure.McpServer.dll"
],
"env": {
"MSSQL_CONNECTIONSTRING": "Server=your_server;Database=your_db;User Id=your_user;Password=your_password;TrustServerCertificate=True;",
"EnableExecuteQuery": "true",
"EnableExecuteStoredProcedure": "true"
}
}
- Save the file and restart Claude Desktop
To use the SQL Server MCP client from a Docker container with Claude Desktop:
- Add the server configuration to the
mcpServers
section in your Claude Desktop configuration:
"mssql": {
"command": "docker",
"args": [
"run",
"--rm",
"-i",
"-e", "MSSQL_CONNECTIONSTRING=Server=your_server;Database=your_db;User Id=your_user;Password=your_password;TrustServerCertificate=True;",
"-e", "EnableExecuteQuery=true",
"-e", "EnableExecuteStoredProcedure=true",
"akashcf/dotnet-mssql-mcp:latest"
]
}
- Save the file and restart Claude Desktop
The server includes a sophisticated type mapping system that converts JSON values to appropriate SQL Server types based on stored procedure parameter metadata:
- Automatic Type Detection: Uses SQL Server's
sys.parameters
metadata as the authoritative source - Rich Type Support: Handles all major SQL Server data types including varchar, nvarchar, int, decimal, datetime, uniqueidentifier, etc.
- Validation: Provides detailed error messages for type mismatches and constraint violations
- Default Values: Supports parameters with default values and optional parameters
- Case-Insensitive: Parameter names are matched case-insensitively
- Flexible Naming: Supports both
@ParameterName
andParameterName
formats - Normalization: Automatic parameter name normalization and validation
- JSON Schema: Generates JSON Schema compatible output for parameter validation
The server implements a multi-layered security approach:
- Tool-Level Security: Individual tools can be enabled/disabled
- Parameter Validation: All inputs are validated against SQL Server metadata
- SQL Injection Protection: Uses parameterized queries throughout
- Connection Security: Supports all SQL Server authentication methods
This project is licensed under the MIT License - see the LICENSE file for details.