read_queryLABSCO SUMMARY
A full hand on the database, not a reporting window.
The same connection that answers questions can insert, update, delete and create tables, and the agent writes the SQL itself after reading the schema. Which account sits in the config file is therefore the decision that matters.
What it is
A Python MCP server for Microsoft SQL Server, adapted from the reference SQLite server. It reads with SELECT, writes with INSERT, UPDATE and DELETE, creates tables, lists the tables that exist and describes a table's schema — so an agent can explore a database it has never seen and then change it. A sixth tool appends findings to a running business-insight memo, so an analysis session leaves something behind besides the transcript.
What you get
- SELECT queries run against the connected database and returned to the client
- INSERT, UPDATE and DELETE over the same connection
- New tables created from the agent's own DDL
- A list of every table in the database
- The schema of any single table, column by column
- A business-insight memo the agent appends findings to as it works
Requirements
A SQL Server the machine can reach, and the Microsoft ODBC driver for SQL Server installed locally. Connection details sit in a config.json beside the server: driver, host, database, and either a username and password or a trusted connection. The account in that file is the account the agent gets.
Setup effort
Build from source — git clone https://github.com/daobataotie/mssql-mcp.git
USE CASES
01
Answering questions against a live database
The agent lists the tables, reads the schema it needs, then writes and runs the SELECT itself.
02
Correcting or loading data without leaving the conversation
INSERT, UPDATE and DELETE run over the same connection, so a fix spotted while reading can be applied on the spot.
03
Standing up a table for a piece of analysis
Create the table, then fill it — a place to stage results the agent produced instead of exporting them somewhere else.
04
Keeping what the analysis found
Findings are appended to a business-insight memo the server maintains, which accumulates across a session.
6 tools
Listed in project docs. Your client may show a different list.
write_querycreate_tablelist_tablesdescribe-tableappend_insightBEFORE YOU INSTALL
What to know first
Tools you get
6 tools named in the project's documentation.
Requires
- Python 3
- ODBC Driver 17 for SQL Server
- A SQL Server database you can log in to
INSTALLATION
Connect your client
Server setup
TERMINAL
git clone https://github.com/daobataotie/mssql-mcp.git
cd mssql-mcp
pip install -r requirements.txtMCP Clients
Claude Desktop, Windsurf
Create config.json in the src folder, next to server.py
Put in your SQL Server address, database name and login:
config.json
{
"database": {
"driver": "ODBC Driver 17 for SQL Server",
"server": "server ip",
"database": "db name",
"username": "<YOUR_USERNAME>",
"password": "<YOUR_PASSWORD>",
"trusted_connection": false
},
"server": {
"name": "mssql-manager",
"version": "0.1.0"
}
}Add the server to your client configuration
<YOUR_PATH_TO_MSSQL_MCP> is the folder you cloned into.
JSON
{
"mcpServers": {
"mssql": {
"command": "python",
"args": [
"<YOUR_PATH_TO_MSSQL_MCP>/src/server.py"
]
}
}
}
