This project provides a Model Context Protocol (MCP) server that allows AI agents to interact with Microsoft SQL Server databases. This version is implemented in Node.js and TypeScript.
- Connects directly to a single MSSQL database instance configured via environment variables.
- Provides tools for:
- Listing tables (
list_tables). - Retrieving table schemas (
get_table_schema). - Performing CRUD operations (
read_table_rows,create_table_records,update_table_records,delete_table_records).
- Listing tables (
- Node.js (v18 or later recommended)
- npm (usually comes with Node.js)
- Access to a Microsoft SQL Server instance.
- Necessary MSSQL ODBC drivers installed on the machine running the server.
- Clone the repository:
git clone <repository-url> cd mcp-server-mssql
- Install dependencies:
npm install
- Environment Variables: Configuration is handled entirely through environment variables.
- Copy
.env.exampleto.env. - Fill in the required MSSQL connection details:
MSSQL_HOSTMSSQL_PORT(defaults to 1433 if not set)MSSQL_USERMSSQL_PASSWORDMSSQL_DATABASE
- Optionally configure other MSSQL options (
MSSQL_DRIVER,MSSQL_ENCRYPT,MSSQL_TRUST_SERVER_CERTIFICATE) and logging (LOG_LEVEL) as described in.env.example.
- Copy
- Build the TypeScript code:
npm run build
- Run the server:
Alternatively, for development:
npm start
npm run dev
- Connect via MCP Client: Configure your MCP client (e.g., Roo) to connect to this server using stdio.
(Detailed descriptions of the tools can be found in the server's ListTools response)
list_tablesget_table_schemaread_table_rowscreate_table_recordsupdate_table_recordsdelete_table_records
(Add details about running tests, linting, etc. later)
[MIT] - (Or your chosen license)