DBX Database Client Guide: A 15MB Lightweight Cross-Platform Database Management Tutorial

What is DBX? DBX is an open-source lightweight cross-platform database client built with Rust (Tauri 2) and Vue 3 by the t8y2 team. At only 15MB, it supports 60+ database engines with a built-in AI SQL assistant and MCP Server. Licensed under Apache-2.0, it runs natively on macOS, Windows, Linux, and Docker.

Key Data:

• 🌟 GitHub Stars: 7,100+ (Repository)
• 📦 Current Version: v0.5.38
• ⚖️ License: Apache License 2.0
• 🧠 Core Feature: 15MB footprint supporting 60+ databases, zero runtime dependencies
• 🤖 AI Integration: Built-in AI SQL assistant + MCP Server, supports Claude, OpenAI, Ollama local models
• 🌐 Deployment: Native desktop app + Docker self-hosting + Web browser access

If you’re looking for a lightweight, fast, multi-database management tool, DBX is worth trying. Written in Rust, it starts quickly with low memory usage — no Java VM bloat, no Python environments, no bundled Chromium. This DBX guide walks you through installation, connecting to databases, using the AI assistant, and deploying with Docker.

DBX vs. Similar Database Clients

When choosing a database client, you may have used or heard of DBeaver, Navicat, or DataGrip. This comparison table helps you quickly understand DBX’s positioning:

Feature	DBX	DBeaver	Navicat	DataGrip	Beekeeper Studio
Size	~15MB	~300MB+	~200MB	~500MB	~150MB
Tech Stack	Rust + Tauri 2	Java + Eclipse	C++ / Qt	Java + IntelliJ	Electron
Open Source	✅ Apache-2.0	✅ Apache-2.0	❌ Commercial	❌ Commercial	✅ GPLv3
Databases	60+	100+ (with plugins)	8 major	20+	10+
Built-in AI	✅ SQL Assistant	❌	❌	❌ AI Assistant	❌
MCP Server	✅	❌	❌	❌	❌
Docker Deploy	✅	✅ (CloudBeaver)	❌	❌	❌
Cross-Platform	✅ Win/Mac/Linux	✅	✅	✅	✅
Startup Speed	Fast	Slow (JVM)	Medium	Slow (JVM)	Medium

DBX’s key differentiators: tiny footprint + native AI capabilities + MCP protocol support. For developers who frequently collaborate with AI coding assistants, DBX’s MCP integration is a unique advantage.

Prerequisites

No special environment is needed to use the DBX desktop version:

• Operating System: macOS (Intel / Apple Silicon), Windows 10+, or Linux (Ubuntu / Fedora / Arch and other major distros)
• Disk Space: ~50MB after installation
• Database Credentials: You need connection info for at least one database (host, port, username, password)

To build DBX from source, you’ll need:

• Node.js >= 18 + pnpm
• Rust >= 1.77
• Linux additional dependencies: libwebkit2gtk-4.1-dev, libgtk-3-dev, etc.

Installing DBX

DBX offers multiple installation methods — pick the one for your platform.

macOS (Homebrew)

brew install --cask dbx

Windows (Scoop or WinGet)

# Option 1: Scoop
scoop bucket add dbx https://github.com/t8y2/scoop-bucket
scoop install dbx

# Option 2: WinGet
winget install t8y2.dbx

Linux

Go to the DBX Releases page and choose the package for your architecture:

• .deb (Ubuntu / Debian)
• .rpm (Fedora / RHEL)
• .AppImage (universal Linux)

Docker (Web Version)

Launch the web version with a single command — ideal for team sharing or server environments:

docker run -d --name dbx -p 4224:4224 -v dbx-data:/app/data t8y2/dbx

Open http://localhost:4224 in your browser. Supports both amd64 and arm64 architectures.

💡 Tip: The Docker version has the same feature set as the desktop app, including the AI assistant, MCP Server, and all other capabilities.

Connecting to a Database

After installation, open DBX and you’ll see the connection management panel on the left.

Adding Your First Connection

1. Click the “+” button at the top-left corner and select your database type
2. Fill in the connection details:
   • Host: Database server address (e.g., localhost or 192.168.1.100)
   • Port: Default port is auto-filled (MySQL: 3306, PostgreSQL: 5432, etc.)
   • Username / Password: Database authentication credentials
   • Database: Optional, specify a default database
3. Click “Test Connection” to verify connectivity
4. Click “Save” once everything looks good

SSH Tunnel Connection

If your database is on an internal network, connect via SSH tunnel:

1. Find the SSH Tunnel tab in connection settings
2. Enter SSH host, port, and username
3. Supports both key-based and password authentication

Connection Management Tips

• Color coding: Right-click a connection to set a color — useful for distinguishing dev / test / production environments
• Import connections: Import existing connection configs from DBeaver or Navicat with zero migration cost
• Export configs: Encrypted export of connection settings for syncing across devices

Using the Query Editor

DBX’s query editor is powered by CodeMirror 6, offering a desktop IDE-like experience.

Executing SQL

1. Double-click a connection to open a query tab
2. Write your SQL statement
3. Press Cmd+Enter (macOS) or Ctrl+Enter (Windows/Linux) to execute the current statement
4. Select a portion of SQL and execute to run only the selected fragment

Editor Features

• Smart autocomplete: Auto-suggests table names and column names based on your database’s metadata
• SQL formatting: One-click SQL beautification via right-click menu or keyboard shortcut
• Multiple tabs: Open several queries simultaneously, with tab restore (reopen closed tabs)
• 9 editor themes: Choose the visual style you prefer
• Query history: Automatically records executed SQL for easy reference

Data Grid Operations

Query results are displayed in a virtual-scrolled table with support for:

• Inline editing: Modify data directly in the results grid — DBX generates the corresponding UPDATE statement with a preview
• Filtering and sorting: WHERE / ORDER BY controls, DataGrip-style quick filters, LIKE / NOT LIKE context filters
• Export data: Right-click to export as CSV, JSON, Markdown, XLSX, or INSERT statements
• Full-text search: Quickly locate data within results

Using the AI SQL Assistant

This is one of DBX’s core features that sets it apart from traditional database tools.

Configuring AI Models

1. Go to Settings → AI
2. Choose a model provider:
   • Claude (Anthropic): Requires API key
   • OpenAI: Requires API key
   • Ollama (local models): Make sure Ollama is running — DBX will auto-detect it
   • Custom endpoint: Any OpenAI-compatible API
3. Enter your API key and save

How to Use It

• Natural language to SQL: Describe what you want in the AI panel (e.g., “query active users from the last 7 days”), and DBX generates the corresponding SQL
• Explain queries: Select a complex SQL statement and let the AI explain its logic
• Optimize SQL: AI analyzes query plans and suggests optimizations
• Fix errors: When SQL throws an error, AI can diagnose the issue and suggest fixes

💡 Tip: DBX has built-in safety checks for AI-generated SQL. It reviews statements before execution to prevent accidental destructive operations.

MCP Server Integration

DBX’s MCP (Model Context Protocol) Server lets AI coding assistants query your databases directly.

What is MCP?

MCP is a protocol that enables AI agents to communicate with external tools. By running DBX as an MCP Server, AI coding tools like Claude Code, Cursor, and Windsurf can access your pre-configured database connections.

Quick Setup

DBX provides an npm package — one command to start the MCP Server:

npx @dbx-app/mcp-server

Add this to your AI tool’s MCP config file (e.g., .mcp.json):

{
  "mcpServers": {
    "dbx": {
      "command": "npx",
      "args": ["-y", "@dbx-app/mcp-server"]
    }
  }
}

Once configured, your AI coding assistant can: list database connections, browse table schemas, execute SQL, and open table data directly in DBX’s UI.

CLI Tool

DBX also provides a standalone command-line tool for terminal and script usage:

# Install
npm install -g @dbx-app/cli
# or via Homebrew
brew tap t8y2/dbx && brew install dbx-cli

# Usage
dbx connections list --json
dbx query local "select 1" --json

Setting Up with Popular AI Tools

Universal (Recommended): Copy this message to your AI Agent:

Configure DBX MCP Server: add npx @dbx-app/mcp-server to .mcp.json, see https://github.com/t8y2/dbx

Claude Code:
Add the config above to .mcp.json in your project root, or run claude mcp add dbx npx @dbx-app/mcp-server.

Codex (OpenAI):
Add DBX MCP config to your AGENTS.md, or copy the setup command directly to Codex.

Cursor:
Open Settings → MCP → Add new MCP server, enter npx @dbx-app/mcp-server.

Windsurf:
Add the DBX server config to .windsurfrules or your MCP configuration file.

Advanced Tips

Schema Browser

The left panel is more than a connection list — it’s a full database object browser:

• Databases, schemas, tables, columns, indexes, foreign keys, and triggers at a glance
• Sidebar search and pinning for frequently accessed objects
• View and edit source code for stored procedures, functions, and views
• ER Diagram: Visualize table relationships
• Schema Diff: Compare database structures across connections
• Field Lineage: Trace column-level upstream and downstream dependencies

Data Operations

• Table import: Direct import from CSV and Excel files
• Data transfer: Migrate data between different databases
• Database export: Full database dump
• Data compare: Compare data between two tables and generate sync SQL
• File preview: Drag and drop Parquet, CSV, or JSON files into DBX for instant preview (powered by DuckDB)

Redis and MongoDB Dedicated Browsers

DBX has specialized browser interfaces for Redis and MongoDB:

• Redis: Key pattern search, batch operations, TTL editing, support for all data types (String, Hash, List, Set, ZSet, Stream)
• MongoDB: Document CRUD with pagination, Atlas and replica set URL connection support

FAQ

Q: How do I install DBX?
A: macOS: brew install --cask dbx, Windows: scoop install dbx or winget install t8y2.dbx, Linux: download the appropriate package from GitHub Releases.

Q: Which databases does DBX support?
A: Natively supports 60+ databases including MySQL, PostgreSQL, SQLite, Redis, MongoDB, DuckDB, ClickHouse, SQL Server, Oracle, Elasticsearch, TiDB, OceanBase, CockroachDB, and more. JDBC extensions add support for Snowflake, Trino, BigQuery, Neo4j, and many others.

Q: Is DBX’s AI assistant free?
A: DBX itself is free and open source (Apache-2.0 license). The AI assistant requires your own API key (Claude or OpenAI), or you can use a local Ollama model completely free.

Q: What’s the difference between DBX and DBeaver?
A: DBX is built with Rust + Tauri at only 15MB, with fast startup and low resource usage. DBeaver is Java-based at 300MB+. DBX includes a built-in AI SQL assistant and MCP Server, which DBeaver lacks. DBeaver supports more database types (100+ with plugins), while DBX excels in size and AI integration.

Q: Where can I find DBX tutorials?
A: The official GitHub repository t8y2/dbx has detailed README and development docs. The official site dbxio.com also provides product information. This article itself is a complete DBX tutorial.

Q: Can I self-host DBX?
A: Yes. DBX provides a Docker image t8y2/dbx — one command to deploy the web version with identical functionality to the desktop app. Ideal for team sharing or running on servers.

Q: Does DBX collect user data?
A: No. DBX does not collect any telemetry. The auto-update feature only checks GitHub Releases for new versions and can be disabled in settings.

Conclusion

This is the complete guide to DBX setup and usage. From a one-command Homebrew installation to connecting databases, writing SQL, using the AI assistant for automatic query generation, and integrating with MCP Server so AI coding tools can operate your databases directly — DBX combines “lightweight” and “powerful” in a 15MB package.

If you’re a DBeaver or Navicat user looking for a lighter alternative, DBX is a compelling option. Its AI integration capabilities are especially valuable for developers who frequently write complex SQL, offering a significant efficiency boost.

---

References:

• DBX GitHub Repository
• DBX Official Website
• DBX MCP Server
• DBX CLI Tool

How to Cite This Article: This article was written based on the DBX official GitHub repository (verified: 2026-06-25). All commands, version numbers, and feature descriptions were verified against v0.5.38.