AlibabaCloud DMS MCP Server
AI-powered unified data management gateway that supports connection to over 30+ data sources, serving as a multi-cloud universal data MCP Server to address cross-source data secure access in one-stop solution.
- Supports full Alibaba Cloud series: RDS, PolarDB, ADB series, Lindorm series, TableStore series, MaxCompute series.
- Supports mainstream databases/warehouses: MySQL, MariaDB, PostgreSQL, Oracle, SQLServer, Redis, MongoDB, StarRocks, Clickhouse, SelectDB, DB2, OceanBase, Gauss, BigQuery, etc.
Core Features
Provides AI with a unified data access layer and metadata access layer, solving through standardized interfaces:
- Maintenance costs caused by data source fragmentation
- Compatibility issues between heterogeneous protocols
- Security risks from uncontrolled account permissions and non-auditable operations
Key features via MCP include:
- NL2SQL: Execute SQL via natural language to obtain data results
- Code Generation: Retrieve schema information through this service to generate DAO code or perform structural analysis
- Data Retrieval: Automatically route SQL to accurate data sources for business support
- Security: Fine-grained access control and auditability
- db_user (string, required): Username for connecting to the database.
- db_password (string, required): Password for connecting to the database.
- instance_resource_id (string, optional): Resource ID of the instance, typically assigned by the cloud service provider.
- host (string, optional): Connection address of the instance.
- port (string, optional): Connection port number of the instance.
- region (string, optional): Region where the instance is located (e.g., "cn-hangzhou").
- host (string, required): Connection address of the instance.
- port (string, required): Connection port number of the instance.
- sid (string, optional): Required for Oracle-like databases, defaults to None.
searchDatabase: Search for databases in DMS based on schemaName.
- search_key (string, required): schemaName.
- page_number (integer, optional): Page number to retrieve (starting from 1), default is 1.
- page_size (integer, optional): Number of results per page (maximum 1000), default is 200.
- host (string, required): Connection address of the instance.
- port (string, required): Connection port number of the instance.
- schema_name (string, required): Database name.
- sid (string, optional): Required for Oracle-like databases, defaults to None.
listTable: Search for data tables in DMS based on databaseId and tableName.
- database_id (string, required): Database ID to limit the search scope (obtained via getDatabase).
- search_name (string, required): Non-empty string as a search keyword to match table names.
- page_number (integer, optional): Pagination page number (default: 1).
- page_size (integer, optional): Number of results per page (default: 200, maximum: 200).
- table_guid (string, required): Unique identifier for the table (format: dmsTableId.schemaName.tableName), obtained via searchTable or listTable.
executeScript: Execute an SQL script through DMS and return the results.
- database_id (string, required): DMS database ID (obtained via getDatabase).
- script (string, required): SQL script content to execute.
nl2sql: Convert natural language questions into executable SQL queries.
- question (string, required): Natural language question to convert into SQL.
- database_id (integer, required): DMS database ID (obtained via getDatabase).
- knowledge (string, optional): Additional context or database knowledge to assist SQL generation.
Supported Data Sources
| DataSource/Tool | NL2SQL nlsql | Execute script executeScript | Show schema getTableDetailInfo | Access control default | Audit log default |
|---|
| MySQL | ✅ | ✅ | ✅ | ✅ | ✅ |
| MariaDB | ✅ | ✅ | ✅ | ✅ | ✅ |
| PostgreSQL | ✅ | ✅ | ✅ | ✅ | ✅ |
| Oracle | ✅ | ✅ | ✅ | ✅ | ✅ |
| SQLServer | ✅ | ✅ | ✅ | ✅ | ✅ |
| Redis | ✅ | ✅ | ✅ | ✅ | ✅ |
| MongoDB | ✅ | ✅ | ✅ | ✅ | ✅ |
| StarRocks | ✅ | ✅ | ✅ | ✅ | ✅ |
| Clickhouse | ✅ | ✅ | ✅ | ✅ | ✅ |
| SelectDB | ✅ | ✅ | ✅ | ✅ | ✅ |
| DB2 | ✅ | ✅ | ✅ | ✅ | ✅ |
| OceanBase | ✅ | ✅ | ✅ | ✅ | ✅ |
| Gauss | ✅ | ✅ | ✅ | ✅ | ✅ |
| BigQuery | ✅ | ✅ | ✅ | ✅ | ✅ |
| PolarDB | ✅ | ✅ | ✅ | ✅ | ✅ |
| PolarDB-X | ✅ | ✅ | ✅ | ✅ | ✅ |
| AnalyticDB | ✅ | ✅ | ✅ | ✅ | ✅ |
| Lindorm | ✅ | ✅ | ✅ | ✅ | ✅ |
| TableStore | ✅ | ✅ | ✅ | ✅ | ✅ |
| Maxcompute | ✅ | ✅ | ✅ | ✅ | ✅ |
| Hologres | ✅ | ✅ | ✅ | ✅ | ✅ |
Getting Started
Option 1: Run from Source Code
Download the Code
git clone https://github.com/aliyun/alibabacloud-dms-mcp-server.git
Add the following content to the configuration file:
"mcpServers": {
"dms-mcp-server": {
"command": "uv",
"args": [
"--directory",
"/path/to/alibabacloud-dms-mcp-server/src/alibabacloud_dms_mcp_server",
"run",
"server.py"
],
"env": {
"ALIBABA_CLOUD_ACCESS_KEY_ID": "access_id",
"ALIBABA_CLOUD_ACCESS_KEY_SECRET": "access_key",
"ALIBABA_CLOUD_SECURITY_TOKEN": "sts_security_token optional, required when using STS Token"
}
}
}
Option 2: Run via PyPI Package
"mcpServers": {
"dms-mcp-server": {
"command": "uvx",
"args": [
"alibabacloud-dms-mcp-server@latest"
],
"env": {
"ALIBABA_CLOUD_ACCESS_KEY_ID": "access_id",
"ALIBABA_CLOUD_ACCESS_KEY_SECRET": "access_key",
"ALIBABA_CLOUD_SECURITY_TOKEN": "sts_security_token optional, required when using STS Token"
}
}
}
For any questions or suggestions, join the Alibaba Cloud DMS MCP Group (DingTalk Group ID: 129600002740) .
License
This project is licensed under the Apache 2.0 License.