Show HN: MCP server to connect LLM agents to any database

16 hours ago 1

Turbular is an open-source Model Context Protocol (MCP) server that enables seamless database connectivity for Language Models (LLMs). It provides a unified API interface to interact with various database types, making it perfect for AI applications that need to work with multiple data sources.

🔌 Multi-Database Support: Connect to various database types through a single API
🔄 Schema Normalization: Automatically normalize database schemas to correct naming conventions for LLM compatibility
🔒 Secure Connections: Support for SSL and various authentication methods
🚀 High Performance: Optimizes your LLM generated queries
📝 Query Transformation: Let LLM generate queries against normalized layouts and transform them into their unnormalized form
🐳 Docker Support: Easy deployment with Docker and Docker Compose

Using Docker (Recommended)

Clone the repository:

git clone https://github.com/raeudigerRaeffi/turbular.git cd turbular
Start the development environment:

docker-compose -f docker-compose.dev.yml up --build
Test the connection:

./scripts/test_connection.py

Install Python 3.11 or higher
Install dependencies:

pip install -r requirements.txt
Run the server:

uvicorn app.main:app --reload

Retrieve the schema of a connected database for your LLM agent.

Parameters:

db_info: Database connection arguments
return_normalize_schema (optional): Return schema in LLM-friendly format

Optimizes query and then execute SQL queries on the connected database.

Parameters:

db_info: Database connection arguments
query: SQL query string
normalized_query: Boolean indicating if query is normalized
max_rows: Maximum number of rows to return
autocommit: Boolean for autocommit mode

POST /upload-bigquery-key

Upload a BigQuery service account key file.

Parameters:

project_id: BigQuery project ID
key_file: JSON key file

Upload a SQLite database file.

Parameters:

database_name: Name to identify the database
db_file: SQLite database file (.db or .sqlite)

Verify if the API is running.

Get a list of all supported database types.

Fork and clone the repository
Create a development environment:

docker-compose -f docker-compose.dev.yml up --build
The development server includes:
- FastAPI server with hot reload
- PostgreSQL test database
- Pre-configured test data
Access the API documentation:
- Swagger UI: http://localhost:8000/docs
- ReDoc: http://localhost:8000/redoc

We welcome contributions! Here's how you can help:

Check out our contribution guidelines
Look for open issues
Submit pull requests with improvements
Help with documentation
Share your feedback

Follow PEP 8 style guide
Write tests for new features
Update documentation as needed
Use meaningful commit messages

Add more testing, formatting and commit hooks
Add SSH support for database connection
Add APIs as datasources using steampipe
Enable local schema saving for databases to which the server has already connected
Add more datasources (snowflake, mongodb, excel, etc.)

Run the test suite:

For development tests with the included PostgreSQL:

./scripts/test_connection.py

connection_info = { "database_type": "PostgreSQL", "username": "user", "password": "password", "host": "localhost", "port": 5432, "database_name": "mydb", "ssl": False }

connection_info = { "database_type": "BigQuery", "path_cred": "/path/to/credentials.json", "project_id": "my-project", "dataset_id": "my_dataset" }

connection_info = { "type": "SQLite", "database_name": "my_database" }

This project is licensed under the MIT License - see the LICENSE file for details.