mcp-db-explorer
A Model Context Protocol (MCP) server that exposes tools for exploring and querying SQLite and PostgreSQL databases. It allows Large Language Models (LLMs) to inspect table structures, perform statistical profiling, generate entity-relationship diagrams, and execute queries safely.
README
Model Context Protocol Database Explorer
A Model Context Protocol (MCP) server that exposes tools for exploring and querying SQLite and PostgreSQL databases. It allows Large Language Models (LLMs) to inspect table structures, perform statistical profiling, generate entity-relationship diagrams, and execute queries safely.
Overview
The Database Explorer MCP Server provides a bridge between LLMs and SQL databases. It features read-only constraints by default, AST-based query safety verification, automatic schema visualization, and execution plan inspection tools.
Architecture
The diagram below outlines the flow of a client request to the database via the MCP server:
graph TD
Client[MCP Client] -->|JSON-RPC| Server[MCP Database Explorer Server]
Server -->|Parse & Safety Check| Safety[Safety Validator]
Safety -->|AST Verification| Allowed{Safe?}
Allowed -->|Yes| Database[(Database: SQLite / PostgreSQL)]
Allowed -->|No| Block[Block Query & Return Error]
Database -->|Query Results| Server
Server -->|Response| Client
Project Structure
mcp-db-explorer/
├── .github/
│ └── workflows/
│ └── ci.yml # GitHub Actions CI workflow
├── mcp_db_explorer/
│ ├── __init__.py
│ ├── database.py # Database connection and querying logic
│ ├── prompts.py # Prompt templates (e.g. NL-to-SQL)
│ ├── safety.py # AST-based SQL safety validation
│ └── server.py # MCP Server definition and CLI entrypoint
├── tests/
│ ├── test_database.py # Tests for database connections and queries
│ ├── test_safety.py # Tests for AST query validation
│ └── test_server.py # Tests for MCP server tool schemas
├── pyproject.toml # Project dependency configuration
└── README.md # Project documentation
Setup and Installation
Prerequisites
Ensure you have Python 3.10 or later installed on your system.
1. Clone the Repository
git clone https://github.com/arman1o1/mcp-db-explorer
cd mcp-db-explorer
2. Setup and Install Dependencies
Option A: Using uv (Recommended)
This project is configured with uv. To install dependencies and set up the virtual environment, run:
uv sync
Option B: Using pip and venv
You can set up a virtual environment and install dependencies manually:
python -m venv .venv
source .venv/bin/activate # On Windows: .venv\Scripts\activate
pip install -e .
Running the Application
Running with uv
To run the MCP server, use the uv run command:
uv run mcp-db-explorer --db-type sqlite --connection-string my_database.db
Configuration Options
The server accepts several options via command-line arguments or environment variables:
| Argument | Environment Variable | Description |
|---|---|---|
--db-type |
DB_TYPE or MCP_DB_TYPE |
Default database engine type (sqlite, postgres, postgresql) |
--connection-string |
DATABASE_URL or MCP_DB_CONNECTION_STRING |
Database connection string (path for SQLite, URI for PostgreSQL) |
--allow-writes |
ALLOW_WRITES |
Enable write operations (DML/DDL queries). Disabled by default. |
--allowed-dir |
ALLOWED_DATABASE_DIR |
Restricts SQLite database paths to be inside this directory. |
Client Integration
To use this server with an MCP client (such as Claude Desktop), add the server configuration to your client configuration file.
Claude Desktop Configuration
Open your claude_desktop_config.json:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS:
~/Library/Application Support/Claude/claude_desktop_config.json
Add the following configuration (use absolute paths):
{
"mcpServers": {
"mcp-db-explorer": {
"command": "uv",
"args": [
"--directory",
"C:\\absolute\\path\\to\\mcp-db-explorer",
"run",
"mcp-db-explorer",
"--db-type",
"sqlite",
"--connection-string",
"C:\\absolute\\path\\to\\my_database.db"
]
}
}
}
Replace the paths with the correct ones for your local setup.
Testing
To run the test suite, run:
uv run pytest
License
This project is licensed under the MIT License - see the LICENSE file for details.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。