dbx is a roughly 25 MB cross-platform database client written in Rust that connects to more than 90 databases and ships as a desktop application, a Docker container, a CLI, and a browser-based web version, aimed at developers and teams who want a single lightweight tool for MySQL, PostgreSQL, SQLite, Redis, MongoDB, DuckDB, SQL Server, Dameng, and many others.

What it is

dbx is an open-source database client and management tool distributed under the Apache-2.0 licence. It runs natively on macOS, Windows, and Linux, and it can also be self-hosted through Docker or accessed through a web interface, with the same connections and feature set across all three delivery forms. The project is written in Rust and covers a very wide driver surface: native support reaches MySQL, PostgreSQL, SQLite, Cloudflare D1, Redis, MongoDB, DuckDB, ClickHouse, SQL Server, Oracle, Elasticsearch, Easysearch, Meilisearch, Qdrant, Milvus, Weaviate, MariaDB, TiDB, OceanBase, openGauss, GaussDB, KWDB, KingbaseES, Vastbase, GoldenDB, Doris, SelectDB, StarRocks, Manticore Search, Redshift, DM, TDengine, XuguDB, CockroachDB, Access, HighGo, UXDB, and Dolt, with agent-based profiles extending that list to H2, Snowflake, Trino, PrestoSQL, Hive, DB2, Informix, Neo4j, Cassandra, BigQuery, Cloud Spanner, Kylin, SunDB, JDBCX, and custom JDBC connections.

The concrete problem it addresses is the weight and licensing model of existing database GUIs. The README states that DBeaver requires a Java runtime and that TablePlus is freemium, and it positions dbx as an alternative that carries no Java JRE, no Python virtual environment, and no bundled Chromium. dbx ships as a single small binary, so a user downloads it, installs it, and connects without provisioning a runtime first. That matters in environments where installing a JVM or a browser engine is inconvenient or prohibited.

Key capabilities

• Connects to 90+ databases from one ~25 MB application, including MySQL, PostgreSQL, SQLite, Redis, MongoDB, DuckDB, ClickHouse, SQL Server, Oracle, Elasticsearch, and Dameng.
• Built-in AI SQL assistant that turns plain-language descriptions into SQL, explains queries, optimizes SQL, and fixes errors, working with Claude, OpenAI, local models via Ollama, or any OpenAI-compatible endpoint.
• Built-in safety checks that review AI-generated SQL before it executes.
• MCP Server support, letting Claude Code, Cursor, Windsurf, and other AI coding agents query databases through existing connections.
• Query editor built on CodeMirror 6 with SQL syntax highlighting, metadata-aware autocomplete, Cmd+Enter execution, selected-SQL execution, diagnostics, SQL formatting, and 9 editor themes.
• Virtual-scrolled data grid for large result sets, with inline editing and SQL preview before changes are applied.
• Message queue administration for Pulsar, Kafka, and RocketMQ.

Who uses it and how

• Individual developers on macOS, Windows, or Linux who want one native binary instead of a JVM-based client plus a separate SQL editor.
• Teams that self-host dbx through Docker so several people share the same database connections and configuration.
• Users in browser-only or locked-down environments who need the web version, which exposes the same feature set and connections as the desktop build.
• AI-assisted development workflows, where Claude Code, Cursor, or Windsurf reach databases over MCP without a separate configuration per tool.
• Multi-engine teams that work across relational, document, search, vector, and message-queue systems, since Pulsar, Kafka, and RocketMQ admin sits alongside the SQL and NoSQL drivers.

Getting started

dbx is obtained as a single small binary from the project homepage at https://dbxio.com, or self-hosted with Docker for team access. No runtime, JVM, or additional dependency is required.

How it compares

The README contrasts dbx directly with DBeaver, which requires Java, and with TablePlus, which is freemium, presenting dbx as an Apache-2.0 alternative that runs with nothing extra installed. On data ownership and cost model, the distinction is that dbx is open source and self-hostable through Docker, so connections and credentials stay under the operator's control rather than inside a commercial licence tier. Those two named tools are the only comparable products the facts supply.

When to use it — and when not to

A self-hoster must operate the Docker container, which is the only server-side deployment path the README describes. Users who need vendor-specific deep tooling for one engine, or who want a commercial support contract, are better served elsewhere. The project also carries 1406 open issues against 19942 stars and 1935 forks, so the breadth of 90+ drivers comes with a large and busy issue tracker that a prospective user should review before standardising on it.