DBHub

What is DBHub about?

DBHub acts as a bridge between various relational databases and tools that understand the Model Context Protocol (MCP). It translates MCP requests into native database operations, enabling AI assistants such as Claude Desktop, Claude Code, Cursor, and other MCP clients to browse schemas, run queries, and retrieve metadata without writing custom connectors.

How to use DBHub?

1. Install – the simplest method is via npx:

   npx @bytebase/dbhub --transport http --port 8080 --dsn "postgres://user:password@localhost:5432/dbname?sslmode=disable"
   

   Replace the DSN with the connection string for MySQL, MariaDB, SQL Server, SQLite, or PostgreSQL.
2. Choose transport – stdio for direct integration with desktop AI tools, or http for network‑based clients (Cursor, MCP Inspector, etc.).
3. Run in demo mode (no external database required):

   npx @bytebase/dbhub --transport http --port 8080 --demo
   

4. Configure read‑only mode if you only need safe exploration:

   npx @bytebase/dbhub --readonly --dsn "postgres://..."
   

5. Add to Cursor by opening the provided deep‑link URL, which pre‑fills the configuration.

Key Features of DBHub

• Multi‑database support: PostgreSQL, MySQL, MariaDB, SQL Server, SQLite.
• Full MCP resource matrix covering schemas, tables, indexes, procedures, and execution tools.
• SQL execution (execute_sql) plus AI‑driven prompts (generate_sql, explain_db).
• Read‑only mode for production safety.
• SSL options (sslmode=disable|require).
• SSH tunnel support (host, port, user, password, private key, passphrase, or SSH config file).
• Docker and npm deployment with a single command.
• Demo mode with a pre‑loaded employee SQLite database for quick testing.

Use Cases of DBHub

• AI‑assisted data exploration – let Claude or Cursor browse schema, generate queries, and get explanations.
• Unified data access layer for internal tools that need to switch between MySQL, Postgres, etc., without code changes.
• Secure read‑only dashboards – expose database metadata via MCP while preventing writes.
• Development & testing – spin up a temporary DBHub instance in Docker or with npx to prototype AI‑driven queries.
• Remote database access – connect through SSH tunnels to databases behind firewalls.

FAQ from DBHub

• Do I need to install a separate server? No. The server runs as a Node.js process launched via npx or Docker.
• Can I use DBHub with SQLite files on my local machine? Yes, use a DSN like sqlite:///path/to/file.db.
• What happens if I send a write‑query in read‑only mode? The request is rejected; only SELECT‑type statements are allowed.
• Is there a limit on the size of the database? DBHub does not impose its own limit; it relies on the underlying database engine.
• How do I secure the HTTP endpoint? Run behind a reverse proxy with TLS, or use the stdio transport for local AI tools.

[!NOTE]
Brought to you by Bytebase, open-source database DevSecOps platform.

            +------------------+    +--------------+    +------------------+
            |                  |    |              |    |                  |
            |                  |    |              |    |                  |
            |  Claude Desktop  +--->+              +--->+    PostgreSQL    |
            |                  |    |              |    |                  |
            |  Claude Code     +--->+              +--->+    SQL Server    |
            |                  |    |              |    |                  |
            |  Cursor          +--->+    DBHub     +--->+    SQLite        |
            |                  |    |              |    |                  |
            |  VS Code         +--->+              +--->+    MySQL         |
            |                  |    |              |    |                  |
            |  Other Clients   +--->+              +--->+    MariaDB       |
            |                  |    |              |    |                  |
            |                  |    |              |    |                  |
            +------------------+    +--------------+    +------------------+
                 MCP Clients           MCP Server             Databases

DBHub is a Universal Database MCP Server implementing the Model Context Protocol (MCP) server interface. This gateway allows MCP-compatible clients to connect to and explore different databases:

• Universal Gateway: Single interface for PostgreSQL, MySQL, MariaDB, SQL Server, and SQLite
• Secure Access: Read-only mode, SSH tunneling, and SSL/TLS encryption support
• Multi-Database: Connect to multiple databases simultaneously with TOML configuration
• Production-Ready: Row limiting, lock timeout control, and connection pooling
• MCP Native: Full implementation of Model Context Protocol with resources, tools, and prompts

Supported Databases

PostgreSQL, MySQL, SQL Server, MariaDB, and SQLite.

MCP Components

DBHub implements MCP Resources, Tools, and Prompts for database operations:

• Resources: Database schema exploration (schemas, tables, indexes, procedures)
• Tools: SQL execution with transaction support
• Prompts: AI-assisted SQL generation and database explanation

Installation

See the full Installation Guide for detailed instructions.

Quick Start

Docker:

docker run --rm --init \
   --name dbhub \
   --publish 8080:8080 \
   bytebase/dbhub \
   --transport http \
   --port 8080 \
   --dsn "postgres://user:password@localhost:5432/dbname?sslmode=disable"

NPM:

npx @bytebase/dbhub --transport http --port 8080 --dsn "postgres://user:password@localhost:5432/dbname?sslmode=disable"

Demo Mode:

npx @bytebase/dbhub --transport http --port 8080 --demo

See Server Options for all available parameters.

Multi-Database Setup

Connect to multiple databases simultaneously using TOML configuration files. Perfect for managing production, staging, and development databases from a single DBHub instance.

See Multi-Database Configuration for complete setup instructions.

Development

1. Install dependencies:

   pnpm install
   

2. Run in development mode:

   pnpm dev
   

3. Build for production:

   pnpm build
   pnpm start --transport stdio --dsn "postgres://user:password@localhost:5432/dbname?sslmode=disable"
   

Testing

See TESTING.md.

Debug

See Debug.

Contributors

Star History