mssql

This MCP server provides AI agents with a dedicated interface to interact with a Microsoft SQL Server database. Through the available tools, you can discover database schemas, list tables and schemas, and execute arbitrary SQL queries directly from the MCP layer. The server uses a .NET-based implementation and Akka.NET for coordination, ensuring robust error handling and clean formatting of results for AI consumption. The primary tools exposed are: execute_sql to run any SQL statement, list_tables to retrieve table metadata including schema, name, type, and row count, and list_schemas to enumerate available schemas/databases in the SQL Server instance. To use it, connect via a Docker container (or your preferred deployment method) and supply a valid MSSQL_CONNECTION_STRING so the MCP server can authenticate against your database. The results are formatted into readable tables and error messages are actionable to help troubleshoot database access issues.

Prerequisites: Docker installed on your system, access to the network to pull the mssql-mcp Docker image, and a valid MSSQL server you can connect to.

1. Install Docker:

• Windows: Docker Desktop
• macOS: Docker Desktop
• Linux: follow your distro's guide to install Docker Engine

2. Pull and Run the MCP server via Docker:

# Run the MCP server container with your MSSQL connection string (example)
docker run -it --rm \
  -e MSSQL_CONNECTION_STRING="Server=host.docker.internal,1533;Database=MyDb;User Id=myUser;Password=myPassword;TrustServerCertificate=true;" \
  mssql-mcp:latest

3. Alternatively, build your own image if you have a custom MSSQL setup:

# If you have a Dockerfile in the repo
git clone https://github.com/Aaronontheweb/mssql-mcp.git
cd mssql-mcp
# Build image locally
docker build -t mssql-mcp:local .

4. Validate connectivity:

• Ensure MSSQL_CONNECTION_STRING is correct and that the SQL Server is reachable from your Docker host.
• Check container logs for startup messages and any connectivity errors.

5. Optional: integrate with your MCP client (Cursor or Claude Desk-top) using the provided configuration examples in the README.

Notes and tips:

• Always supply a valid MSSQL_CONNECTION_STRING; incorrect credentials or network access will result in startup or runtime errors.
• If using Windows Authentication, ensure the connection string is formatted accordingly and that the container can reach the Windows-hosted SQL Server.
• When running in Docker, use host.docker.internal (or an appropriate host address) to reach a locally hosted SQL Server from inside the container.
• For production, consider enabling TLS/Encrypt in the connection string and managing secrets securely (do not hard-code credentials in your config).
• The MCP server supports standard SQL statements (SELECT/INSERT/UPDATE/DELETE) and DDL as needed by your agents; errors are surfaced with actionable messages for easier debugging.
• If you require multiple MCP servers, repeat the mcpServers entry with distinct names and separate connection strings.