MySQL MCP Server
MySQL MCP 服务器
用于与 MySQL 数据库交互的 MCP 服务器。
该服务器支持执行只读查询(query)和最终回滚的写查询(test_execute)。
设置
环境变量
将以下环境变量添加到~/.mcp/.env :
MYSQL_HOST=host.docker.internal # Hostname to access host services from Docker container
MYSQL_PORT=3306
MYSQL_USER=root
MYSQL_PASSWORD=your_password注意:
host.docker.internal是一个特殊的 DNS 名称,用于从 Docker 容器访问主机服务。连接到主机上运行的 MySQL 服务器时请使用此设置。如果连接到其他 MySQL 服务器,请更改为相应的主机名。
mcp.json 配置
{
"mcpServers": {
"mysql": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"--add-host=host.docker.internal:host-gateway",
"--env-file",
"/Users/username/.mcp/.env",
"ghcr.io/xiangma9712/mcp/mysql"
]
}
}
}Related MCP server: MCP MySQL Server
用法
启动服务器
docker run -i --rm --add-host=host.docker.internal:host-gateway --env-file ~/.mcp/.env ghcr.io/xiangma9712/mcp/mysql注意:如果您使用的是 OrbStack,则会自动支持
host.docker.internal,因此可以省略--add-host选项。虽然 Docker Desktop 通常也自动支持此功能,但为了获得更好的可靠性,建议添加--add-host选项。
可用命令
1.执行只读查询
{
"type": "query",
"payload": {
"sql": "SELECT * FROM your_table"
}
}回复:
{
"success": true,
"data": [
{
"id": 1,
"name": "example"
}
]
}2.测试查询执行
{
"type": "test_execute",
"payload": {
"sql": "UPDATE your_table SET name = 'updated' WHERE id = 1"
}
}回复:
{
"success": true,
"data": "The UPDATE SQL query can be executed."
}3. 列出表格
{
"type": "list_tables"
}回复:
{
"success": true,
"data": ["table1", "table2", "table3"]
}4.描述表
{
"type": "describe_table",
"payload": {
"table": "your_table"
}
}回复:
{
"success": true,
"data": [
{
"Field": "id",
"Type": "int(11)",
"Null": "NO",
"Key": "PRI",
"Default": null,
"Extra": ""
},
{
"Field": "name",
"Type": "varchar(255)",
"Null": "YES",
"Key": "",
"Default": null,
"Extra": ""
}
]
}实现细节
使用 TypeScript 实现
使用 mysql2 包
作为 Docker 容器运行
通过标准输入接受 JSON 命令
通过标准输出返回 JSON 响应
使用
host.docker.internal连接到主机 MySQL(兼容 OrbStack 和 Docker Desktop)
安全注意事项
使用环境变量进行敏感信息管理
预防 SQL 注入是实施者的责任
生产使用需要正确的网络配置
连接主机服务时需要适当的防火墙设置
Available Tools
4 toolsdescribe_tableC
Retrieves column information for a table.
Inputs: tableName (required).
list_tablesB
Retrieves a list of tables in the database.
queryB
Executes a read-only SQL query.
Inputs: sql (required).
test_executeB
Checks if an SQL query can be executed and rolls back afterward.
Inputs: sql (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
Each tool has a clearly distinct purpose: describe_table for column metadata, list_tables for table enumeration, query for read-only SQL execution, and test_execute for query validation with rollback. There is no overlap in functionality, making tool selection straightforward for an agent.
The naming follows a consistent verb_noun pattern (describe_table, list_tables, query, test_execute), with 'query' as a minor deviation as it lacks a noun suffix. Overall, the pattern is predictable and readable, though not perfectly uniform.
With 4 tools, the count is reasonable for a database server, covering essential operations like listing tables, describing schema, querying, and testing queries. It is slightly lean but well-scoped, lacking only advanced features like write operations or transaction management.
The toolset covers basic read and validation operations but has notable gaps: there are no tools for data manipulation (e.g., insert, update, delete), schema modification (e.g., create_table), or transaction control. This limits the server to read-only and diagnostic tasks, which may cause agent failures for write workflows.
Maintenance
Related MCP Connectors
Work with your database from AI tools: ask, export, approved changes, and scheduled jobs.
PostgreSQL, MySQL, and SQL Server in one session. 26 read-only MCP tools for AI agents.
Hosted MySQL MCP server for Claude, ChatGPT and Cursor, with each action switched per account.
Text-to-SQL MCP server: read-only queries on PostgreSQL, MySQL and SQL Server with your real schema
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceEnables secure interaction with MySQL databases, allowing AI assistants to list tables, read data, and execute SQL queries through a controlled interface.-
- AlicenseNot gradedqualityDmaintenanceEnables interaction with MySQL databases through MCP, supporting query execution, table operations (insert, update, delete), and schema inspection for natural language database management.ISC
- AlicenseNot gradedqualityDmaintenanceEnables AI models to interact with MySQL databases through standardized operations including querying, executing commands, listing tables, and describing table structures with secure prepared statement support.MIT
- AlicenseAqualityDmaintenanceEnables querying local MySQL databases and managing Docker containers (listing, logs) through natural language via the Model Context Protocol.455 npmISC