ChatSQL MCP Server
Enables natural language queries to PostgreSQL by converting questions to SQL using Google Gemini and executing via FastMCP.
README
💬 ChatSQL — Natural Language to SQL using Gemini & FastMCP
ChatSQL is a command-line tool that lets you query a PostgreSQL database using natural language. It uses Google Gemini to convert your questions into SQL, and FastMCP to execute them against your database.
🚀 Tech Stack
- Python 3.11+
- PostgreSQL (via
psycopg2) - FastMCP (Modular Composable Protocol)
- Google Gemini (for Natural Language to SQL conversion)
- Rich (for pretty command-line interface output)
- python-dotenv (for environment variable configuration)
🧠 Use Case
You ask a question in plain English, and ChatSQL generates and executes the SQL, then displays the results.
Input: "Get all employees earning more than 50000"
Generated SQL:
SELECT * FROM employees WHERE salary > 50000;
Output:
+----+---------+--------+------------+-----------+------------+
| id | name | salary | phone_no | address | department |
+----+---------+--------+------------+-----------+------------+
| 1 | Alice | 75000 | 9876543210 | New York | HR |
| 2 | Bob | 65000 | 1234567890 | Chicago | Finance |
+----+---------+--------+------------+-----------+------------+
⚙️ Setup Instructions
✅ 1. Clone the Repository & Configure Environment
> git clone https://github.com/thomasjerard/ChatSQL.git
> cd chatSQL
✅ 2. Prerequisites
- Bash shell is required (default on macOS/Linux).
- Python 3.11+ must be installed and accessible via your terminal. If not installed,
For macOS:
brew install python@3.11
For Linux:
sudo apt install python3.11 python3.11-venv
✅ 3. Create a .env file in the root folder with the following content:
DB_NAME=chattosql
DB_USER=postgres
DB_PASS=yourpassword
DB_HOST=localhost
DB_PORT=5432
GEMINI_API_KEY=your-gemini-api-key
✅ 3. Run the Application
You’ll need two separate terminals:
🖥️ Terminal 1 — Start the Server
./run_server.sh
💬Terminal 2 — Start the Client
./run_client.sh
Once the client starts, you'll see a prompt like:
💬 ChatSQL (Gemini + FastMCP) 🗨️ Ask (NL):
🔧 What Happens Internally
-
Each script (run_server.sh / run_client.sh) performs the following:
-
Deletes any existing virtual environment and creates a fresh one.
-
Loads environment variables from the .env file.
-
Installs required Python dependencies from requirements.txt.
-
Sets up PostgreSQL:
-
Creates the database (chattosql)
-
Adds the user (postgres)
-
Creates a sample employees table
-
Starts the FastMCP server or the ChatSQL CLI client, depending on the script.
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。