DBX

1. Project/Product Overview

• core highlight sentence: * "15MB Rust binary, instead of 450MB DBeaver JDK-also allows AI to write SQL for you directly. 」

2. What does it mostly do?

DBX is a full-featured database desktop management and web collaboration tool with core capabilities covering the full lifecycle of database management:

2.1 Multi-database connection management (60 types)

Supports native connections to almost all mainstream and domestic databases:

2.2 Intelligent SQL Editor

Professional editor based on CodeMirror 6, supports:

-SQL syntax highlighting and metadata-aware auto-completion

-'Cmd Enter' execution, select SQL execution

-SQL formatting and diagnostic tips

-9 editor themes

-Query history, save SQL code snippet, restore tab

-'. SQL' file directly executed

2.3 AI SQL Assistant (built-in)

-Natural language→SQL: After selecting a table, use natural language to describe the requirements and AI directly generate SQL statements.

-Query Interpretation: provides natural language interpretation of complex SQL statements.

-SQL Optimization: Analyze and optimize slow queries

-Bug fix: Automatically diagnose SQL errors and suggest fixes

-Security Check: SQL statements generated by the AI are executed after the built-in security review.

-Model support:Claude, OpenAI(GPT series), Ollama local model, any OpenAI compatible endpoint

2.4 data browsing and editing (Data Grid)

-Virtual scrolling to handle very large result sets

-SQL preview before saving in-line editing

-DataGrip style filter, LIKE / NOT LIKE right-click filter

-Sort, Full Text Search, Pagination

-Column width adjustment, automatic adaptation, row number, zebra pattern

-Export formats: CSV, JSON, Markdown, XLSX, INSERT statement

2.5 Schema Management Tools

-Schema Browser: Database → Schema→ Table → Fields → Index→ Foreign Keys → Triggers, supporting sidebar search and top setting

-Object Browser: allows you to browse procedures, functions, and views by group, and supports source code editing.

-Table Structure Editor: Visually modify fields and indexes (support review changes)

-ER Diagram: visualizes the association between tables

-Schema comparison: compares table structure differences across connections

-Execution Plan: Visual query execution plan (Explain plan)

-Field Consanguinity Analysis: column-level consanguinity tracking

-Database Global Search: quickly search for objects in a large schema.

2.6 Data Manipulation Tool

-Data Import: Import CSV and Excel files

-Data Migration: Cross-database migration (for example, MySQL → PostgreSQL)

-Database Export: complete database dump

-Data Comparison: compares table data and generates synchronization SQL statements

-File Preview: Drag Parquet, CSV, or JSON to preview files (based on DuckDB)

-Import Connection: Import existing connection configurations from a DBeaver or Navicat

2.7 special browser

-Redis: supports key mode search, batch key operations, command executor, TTL editing, and all data types (String, Hash, List, Set, ZSet, and Stream)

-MongoDB: document CRUD, page browsing, Atlas and replica set URL direct link

2.8 Security and Connectivity

-SSH tunnel (key password authentication)

-Database and AI proxy settings

-Automatic reconnection after disconnection

-Dangerous operation confirmation dialog box (anti-misoperation)

-Connection configuration encryption export/import

-Connection color marking

-Driver Store optional JDBC plug-in

2.9 AI Programming Assistant Integration (MCP Server)

This is DBX most differentiated ability-to have AI coding assistants directly manipulate the database:

npx @dbx-app/mcp-server

To configure '.mcp.json':

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

MCP Server provides nine tools:

SQL security mechanism:

-Allow regular writes by default ('INSERT', 'UPDATE', 'DELETE ... WHERE ...')

-'DBX_MCP_ALLOW_WRITES = 0 'can be forced read-only

-Dangerous statements such as 'DROP', 'TRUNCATE', and 'ALTER' are intercepted by default and must be explicitly enabled by 'DBX_MCP_ALLOW_DANGEROUS_ SQL = 1'

2.10 CLI Tools

Stand-alone command-line tools support terminal, scripting, and automated workflows:

npm install -g @dbx-app/cli
# 或 Homebrew
brew tap t8y2/dbx && brew install dbx-cli

dbx connections list --json
dbx query local "select count(*) from users" --json
dbx schema list local --json
dbx context local --tables users,orders  # 打印 Schema 上下文，方便喂给 AI
dbx open local users  # 在 DBX 桌面端打开表

3. Applicable Scenario

3.1 Daily Database Development and Management

DBX is the "Swiss Army Knife" of database developers-a tool that covers the entire process of development, query, debugging, design, and export. Particularly suitable:

-Back-end developers quickly add, delete, check and change data during the development process

-DBA daily inspection, execution plan analysis, structure comparison

-Unified management tool in multi-database environment (no need to switch back and forth between Navicat, Redis Desktop Manager and MongoDB Compass)

3.2 Localization Substitution and Created Environment

DBX has outstanding performance in domestic database support:

-Native support for Dameng DM, OceanBase, openGauss, GaussDB, KingBase, Xugu XuguDB, Vastbase, GoldenDB, YBDB, YashanDB, GBase 8a/8s, OscarDB and other domestic databases

-Open source protocol (Apache-2.0), no authorization and interruption risk

-Lightweight, no external runtime dependency, suitable for intranet/bastion environment

3.3 AI enhanced database workflow

This is the core scenario that distinguishes DBX from traditional database tools:

-AI-assisted SQL writing: business personnel describe requirements in natural language, generate SQL AI, and execute it after technical personnel review.

-AI programming assistant database access: MCP Server allows AI assistants such as Claude Code and Cursor to directly check the database verification logic during coding.

-The schema context is fed to the AI: The 'dbx context' command of the CLI generates a compact schema description as the context of the AI Prompt

3.4 Team Collaboration (Docker Self-Hosting)

-Docker deployment web version, team members through the browser unified access

-One configuration connection, team sharing

-Support for reverse proxy subpath deployment ('DBX_PUBLIC_BASE_PATH =/dbx')

3.5 Lightweight Alternative

-Alternative DBeaver(450MB, JDK required) → DBX(15MB, no runtime required)

-Alternative TablePlus(macOS exclusive, paid) → DBX (full platform, open source)

-Alternative Navicat (commercial license expensive) → DBX (open source free)

4. Not quite the scene

5. Core Competence List

6. Architecture/deployment/integration approach

6.1 Technology Stack

6.2 installation method

macOS:

brew install --cask dbx

Windows:

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

# WinGet
winget install t8y2.dbx

Linux: Download '.deb' / '.rpm'/'from Releases'/'. AppImage' package

Nix/NixOS: support for flake.nix

6.3 Docker self-hosting

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

Docker Compose:

services:
  dbx:
    image: t8y2/dbx
    ports:
      - "4224:4224"
    volumes:
      - dbx-data:/app/data
    restart: unless-stopped
    # 可选：反向代理子路径
    environment:
      - DBX_PUBLIC_BASE_PATH=/dbx
volumes:
  dbx-data:

The browser accesses 'http:// localhost:4224 ', and multi-schema images (amd64/arm64) are provided.

6.4 Development Environment Construction

# 环境要求
# Node.js >= 18, pnpm, Rust >= 1.77

# 桌面端开发
make

# Web 版开发
make dev-web       # 前端
make dev-backend   # 后端

# 快速模式（跳过 DuckDB 编译，加速本地构建）
make dev-fast
make cargo-check-fast
make cargo-test-fast

# 打包
make package
# 安装包在 src-tauri/target/release/bundle/

6.5 integration mode

How to use #7.

7.1 desktop end to get started quickly

1. Install → 'brew install -- cask dbx'(macOS) or download the installation package
2. Add Connection → Click "" in the upper left corner → Select Database Type → Fill in Host/Port/User Name/Password
3. Start Query * → Click Connect → Write Query in SQL Editor → 'Cmd Enter' Execute
4. Browse structure → View tables, fields, indexes, etc. in the schema browser on the left

7.2 AI SQL Assistant Usage Process

1. 选中一张表（左侧 Schema 浏览器）
2. 打开 AI 面板
3. 输入自然语言："查询过去 30 天每个用户的订单总数和总金额"
4. AI 生成 SQL
5. （可选）人工审核 SQL → 安全检查自动执行
6. 点击执行 → 查看结果

7.3 the Docker deployment process

# 1. 拉取镜像
docker pull t8y2/dbx

# 2. 启动容器
docker run -d --name dbx -p 4224:4224 -v dbx-data:/app/data t8y2/dbx

# 3. 浏览器打开 http://localhost:4224

# 4. 配置数据库连接（Web 界面操作，同桌面端）

# 5. 团队成员通过同一地址访问

7.4 MCP AI Programming Assistant Workflow

开发者（在 Claude Code / Cursor 中）：
  "帮我查一下 users 表有多少行"
  → AI 助手调用 dbx_execute_query 获取结果
  → "users 表有 15,432 行"

  "列出 orders 表的所有字段"
  → AI 助手调用 dbx_describe_table
  → "orders 表有 id, user_id, amount, status, created_at 等 8 个字段"

  "写一个查询最近 7 天订单金额总和的 SQL"
  → AI 助手调用 dbx_get_schema_context 获取表结构
  → 生成 SQL → 调用 dbx_execute_query 验证
  → 返回带数据的完整查询结果

7.5 CLI Automation Script Example

# 查看所有连接
dbx connections list --json | jq '.[].name'

# 导出查询结果为 CSV
dbx query prod "select id, name, email from users limit 100" --format csv > users.csv

# 获取 Schema 上下文（喂给 AI 提示词）
dbx context local --tables users,orders,products > schema_context.txt
cat schema_context.txt | codex exec "帮我校验这些表的外键关系是否合理"

# 在 DBX 桌面端快速打开某个表
dbx open local users

7.6 Data Migration Example

操作路径：菜单栏 → 工具 → 数据传输

1. 选择源连接（如 MySQL）
2. 选择目标连接（如 PostgreSQL）
3. 选择要迁移的表
4. 预览映射关系（字段类型自动转换）
5. 执行迁移
6. 查看迁移结果

8. What can I say before sales

8.1 a sentence positioning

• * "DBX is a 15MB open source database management artifact -30 times lighter than DBeaver, does not need JDK, covers 60 kinds of databases, and has a built-in AI assistant to help you write SQL. "**

8.2 customer pain points → solutions

8.3 differentiated selling points vs competing products

vs DBeaver (the most widely available open source database tool):

-DBeaver requires JDK(~ 200MB),DBX only 15MB, no runtime required

-DBeaver JDBC driver dependencies need to be downloaded manually (troublesome in intranet environment),DBX native connection is available out of the box

-DBeaver AI function requires a separate plug-in, DBX native built-in AI assistant

-DBeaver no MCP Server,DBX native support AI programming assistant integration

-DBeaver Community Edition has restrictions on some advanced features (e. g. ER Chart Enterprise Edition only),DBX is all free

-Positioning: DBX is the DBeaver "Lightweight AI Enhancement" alternative

vs TablePlus(macOS Paid Database Tool):

-TablePlus only supports macOS,DBX full platform (macOS / Windows / Linux)

-TablePlus paid ($89/person),DBX open source free

-TablePlus there is no Docker Web Edition, DBX supports team deployment

-DBX native integration TablePlus no AI and MCP

vs Navicat (commercial database management tools):

-Navicat commercial license is expensive ($1,199/person),DBX is open source and free

-Navicat support for domestic databases needs to be upgraded to a paid version, DBX is all free.

-Navicat no AI capability, DBX built-in AI MCP

vs DataGrip(JetBrains IDE):

-DataGrip part of heavyweight IDE, subscription required ($199/year)

-DataGrip is a development environment, DBX is more lightweight and suitable for non-IDE scenarios

-DBX has Docker web version (team shared),DataGrip only desktop version

-DataGrip there is no MCP,DBX allows AI assistants to directly manipulate the database.

vs Redis Desktop Manager / MongoDB Compass (Special Tools):

-DBX a tool unified management of multiple databases, no need to install multiple special tools

-DBX's dedicated browser is comparable to dedicated tools in core capabilities

8.4 Customer Value Story Line (5-Step Approach)

1. Cut in: "What tool does your team use to manage the database? How many people use it? Is it a genuine license? Is it inconvenient for several people to switch several client tools at ordinary times? 」
2. Resonance : "Many teams face the same problem-the DBeaver is too heavy to need Java, the Navicat is too expensive to buy the original version, the TablePlus can only be used on the Mac, and there are no tools available for the domestic database in the letter creation environment. What is even more troublesome is that in the AI era, database tools cannot be linked with AI. 」
3. Presentation:

-Live download installation (5 seconds to complete) → Connect a MySQL/PostgreSQL(30 seconds)

-Display schema browsing, SQL editing and execution, and data tables

-Highlight Moment 1: Select Table → AI Assistant Describes in Chinese → Automatically Generate SQL and Execute

-Highlight Moment 2: Display ER diagram, Schema comparison, field consanguinity

- Highlight moment 3 : show MCP configuration → directly let the AI check the database in Claude Code

1. Advanced : "DBX is not just a desktop tool-after Docker is deployed, the entire team accesses it through a browser, and the connection configuration is managed in a unified way. MCP allows AI programming assistants to check libraries. A 15MB tool covers 90% of DBA's daily operation scenarios. 」
2. End *: "Open Source Free (Apache-2.0), Active Community (QQ Group/WeChat Group/Discord),14,000 + Stars, Recommended by Product Hunt. The domestic database is fully supported and the letter is environmentally friendly. 」

8.5 Role-based Sales Speech

To Development Manager/Technical Leader:

"DBX allows development teams to switch tools less-MySQL, Redis, MongoDB, ClickHouse one tool. The AI assistant lowers the SQL threshold, and newcomers can get started quickly. MCP Server allows AI programming assistants to directly verify database logic, improving development efficiency by at least 30%. 」

For DBA:

"Schema comparison, execution plan, field kinship analysis-these are the core requirements of DBA. ER diagrams make table relationships clear at a glance. Data migration features cross-database transfer. The most important thing is: free and open source, no licensing risk. 」

For the person in charge of operation and maintenance/security:

"Docker deployment is very simple, a command to start. Support SSH tunnel encryption connection. The connection configuration is encrypted and saved. There is a confirmation pop-up window for hazardous operations. The intranet environment is fully available and does not require networking. 」

• For CIO/CTO (Credit Creation Scenario): *

"DBX native support dream, OceanBase, GaussDB, KingBase, virtual valley and other 10 domestic database. Apache-2.0 open source protocol, without any risk of authorization and interruption. 15MB of Rust binary, no supply chain dependency issues. 」

Budget-sensitive small and medium-sized teams:

"Open source is free. One tool replaces the Navicat Redis Desktop Manager MongoDB Compass, saving at least 2-30000 software licensing fees per year. The 15MB download is ready to use, no Java environment is required, and IT support costs are almost zero. 」

9. Frequently Asked Customer Questions

10. PoC Recommendations

Recommended PoC Direction 1: Multi-database Unified Management

Objective: Verify that DBX can replace the team's current database management tools.

Overall time: 3.5 days

Validation Metrics:

-Can replace at least 2 database management tools currently in use

-Daily operation efficiency is not lower than the current tool

-Advanced features (Schema comparison, ER diagram, data migration) meet basic needs

-Team's experience rating for Web version ≥ 4/5

Recommended PoC Direction 2: Enhance Development Workflow AI MCP

Objective: Verify that the AI SQL Assistant and MCP Server of DBX actually improve the development efficiency.

Total time: 3 days

Validation Metrics:

-AI SQL generation accuracy ≥ 80% (for common business query scenarios)

-AI-assisted SQL writing time reduced by ≥ 50%

-MCP allows AI programming assistants to save at least 20% of the "code → check library → modify" cycle time

-Developer's subjective satisfaction ≥ 4/5

Recommended PoC Direction 3: Domestic Database Management in Xinchuang Environment

Objective: Verify the availability and experience of DBX in the trust environment (domestic OS domestic database).

Total time: 3 days

Validation Metrics:

-All target domestic databases can be successfully connected and perform basic operations

-At least on par with the daily experience of existing tools

-No exception for offline operation

-The team is willing to use it as a daily tool

PoC Success Criteria (Go/No-Go)

Go:

-Core databases (at least 3) can be connected and operated normally

-Daily development/DBA operation (query, edit, export, Schema browsing) experience is smooth

-(AI direction) The SQL generated by the AI is highly efficient and has no missing reports in the security check.

-(Xinchuang Direction) Domestic libraries support standards and are available offline

No-Go:

-Core business database connection failure or frequent disconnection

-Data inconsistency in data migration

-User operating habits are too different from existing tools and cannot be adapted

-Security assessment failed (e. g., enterprise security policy does not allow open source tools to connect directly to production libraries)

PoC promotion suggestion: Start with the desktop version and let 2-3 developers/DBAs use it for 1-2 weeks in daily work. If the feedback is positive, then push the Docker version team deployment.

11. Risks and Considerations

12. My Pre-Sales Judgment

• * Recommendation: Strongly recommended (★★★★★)**

Especially suitable for the scene:

-The team uses multiple databases and requires unified management tools

-Government and enterprise customers with demand for credit creation/localization substitution.

-Development modernization teams that need to AI enhance database workflows

-Small and medium-sized teams with limited budgets and sensitive software licensing costs

Core reasons:

1. "lightweight not light energy" the ultimate experience :15MB Rust binary completely solve the database tool "install a client to half an hour" pain point. This is of great value in the localization/intranet scenario-no JDK, no Python, no embedded Chromium, just copy a file.

1. AI MCP's differentiated moat : This is DBX's most unique core competitiveness. AI SQL Assistant lowers the SQL threshold, MCP Server opens up AI Programming Assistant and database-this is a capability that all competitors such as DBeaver, Navicat and TablePlus do not currently have. MCP ecology is rising rapidly, DBX is the first card position, the first-mover advantage is obvious.

1. Domestic database" timely rain ": letter wave, a large number of enterprises from Oracle/MySQL migration to dream/GaussDB/KingBase/OceanBase, but the market for these domestic library support good client tools are very scarce. DBX's native support is the most direct entry point-it can be positioned as the "preferred tool for database management 」.

1. Open Source Free Active Community :Apache-2.0 Protocol Eliminates Authorization Concerns, Active Chinese Community (QQ Group/WeChat Group/LINUX DO) Lowering Usage Threshold. The growth rate of 14,000 Stars proves that the market demand is real.

1. Strong multi-terminal deployment flexibility : desktop version (personal) Docker Web version (team) CLI (automation) MCP(AI integration), covering the whole scene from individual developers to enterprise teams.

Recommended customer profile (prioritized):

1. government-enterprise trust customers : need to manage domestic databases, no suitable client tools → DBX is the best choice
2. Technical team using multiple databases :MySQL Redis MongoDB ClickHouse hybrid architecture → DBX unified management
3. AI friendly modernization team : Claude Code/Cursor MCP → DBX is already being used to complete the database access capability
4. Small and medium-sized teams/individual developers : Can't use Navicat, and don't want to install DBeaver → DBX is the best alternative
5. DBA Team: As an auxiliary tool for daily inspection and data manipulation

Not recommended situations:

-The team only uses one database and has deep binding special tools (for example, only MySQL and all DataGrip users)

-The enterprise security policy strictly limits the scenario where desktop clients are directly connected to the database and only allows the Web terminal of the bastion machine.

-Enterprise-level scenarios with extremely high performance requirements for database tools (such as OLAP clusters that manage hundreds of millions of rows of tables)

-Requires complex BI reports and visual analysis (DBX does not do BI)

• One word recommendation: * "If your team is still paying for database tools, or if you can't find a good client because of Xinchuang migration -- DBX is worth 5 minutes to download. 」

13. REFERENCE

-GitHub repository: https://github.com/t8y2/dbx

-Chinese README:https://github.com/t8y2/dbx/blob/main/README.zh-CN.md

-MCP Server Documentation: https://github.com/t8y2/dbx/blob/main/packages/mcp-server/README.md

-CLI Documentation: https://github.com/t8y2/dbx/blob/main/packages/cli/README.md

-Nix installation instructions: https://github.com/t8y2/dbx/blob/main/README-NIX.md

-Contribution Guide: https://github.com/t8y2/dbx/blob/main/CONTRIBUTING.md

-Security policy: https://github.com/t8y2/dbx/blob/main/SECURITY.md

-License (Apache-2.0):https://github.com/t8y2/dbx/blob/main/LICENSE

-Official website: https://dbxio.com

-Offline driver download: https://dbxio.com/en/drivers

-Product Hunt:https://www.producthunt.com/products/dbx/launches/dbx

-HelloGitHub recommendation: https://hellogithub.com/repository/t8y2/dbx

-Trendshift trends: https://trendshift.io/repositories/26775

-Domestic Mirror (AtomGit):https://atomgit.com/t8y2/dbx

-Scoop Bucket:https://github.com/t8y2/scoop-bucket

-NPM MCP Server:https://www.npmjs.com/package/@dbx-app/mcp-server

-NPM CLI:https://www.npmjs.com/package/@dbx-app/cli

-QQ group: 1087880322

-Discord:https://discord.gg/W7NyVDRt6a

• Analysis Date: 2026-07-02 | Data Aging: GitHub Real-Time Acquisition, README Content Up to Access Date, DBX v0.5.43 *