Raw SQL MCP
<!-- mcp-name: io.github.KifoPL/RawSqlMcp -->Raw SQL MCP is a STDIO MCP server for executing raw SQL against SQL Server, SQLite, PostgreSQL, MySQL, and MariaDB.
IMPORTANT
Use it carefully. The server does not protect you from destructive queries, expensive queries, SQL injection, missing pagination, or unsafe data exposure. Always review and approve queries before running them.

Installation
Run the server with dnx:
dnx RawSqlMcp
For a specific version:
dnx RawSqlMcp@0.0.2
Package: RawSqlMcp on NuGet
Configuration
Databases are configured with environment variables under RawSqlMcp__Databases__.
SQLite:
export RawSqlMcp__Databases__Local__Provider="sqlite"
export RawSqlMcp__Databases__Local__ConnectionString="Data Source=/absolute/path/app.db"
PostgreSQL:
export RawSqlMcp__Databases__Reporting__Provider="postgres"
export RawSqlMcp__Databases__Reporting__ConnectionString="Host=localhost;Port=5432;Database=reporting;Username=postgres;Password=postgres"
MySQL/MariaDB:
export RawSqlMcp__Databases__Shop__Provider="mysql"
export RawSqlMcp__Databases__Shop__ConnectionString="Server=localhost;Port=3306;Database=shop;User ID=mysql;Password=mysql"
SQL Server:
export RawSqlMcp__Databases__Default__Provider="sqlserver"
export RawSqlMcp__Databases__Default__ConnectionString="Server=localhost,1433;Database=master;User Id=sa;Password=Your_password123;TrustServerCertificate=True"
Optional command timeout, in seconds:
export RawSqlMcp__CommandTimeout=30
[!INFO] The legacy
RawSqlMcp__ConnectionStrings__Default="..."format is obsolete. It still works for compatibility and is interpreted as SQL Server.
Usage
Register the server in an MCP client as a STDIO server:
{
"servers": {
"raw-sql-mcp": {
"command": "dnx",
"args": ["RawSqlMcp"],
"env": {
"RawSqlMcp__Databases__Default__Provider": "sqlserver",
"RawSqlMcp__Databases__Default__ConnectionString": "Server=localhost,1433;Database=master;User Id=sa;Password=Your_password123;TrustServerCertificate=True"
}
}
}
}
The server exposes tools for listing configured database names, reading schema metadata, and executing raw, parameterized, and scalar SQL queries.
