Manage multiple databases and message queues with a single lightweight tool
When working across scattered data sources like MySQL, PostgreSQL, Redis, and Kafka, one tool is enough. DBX is a cross-platform database client distributed as a 25MB single binary. It is available for desktop, Docker, CLI, and web, and supports query writing and AI agent integration through a built-in AI assistant and MCP server.
25MB single binary and 100+ databases
DBX is distributed as a single binary of approximately 25MB that runs without Java JRE, Python virtual environments, or Chromium bundling. It supports macOS, Windows, and Linux desktops, as well as Docker and web versions, allowing the same features to be used in various environments. It can connect to over 100 databases, including MySQL, PostgreSQL, SQLite, Redis, MongoDB, DuckDB, ClickHouse, SQL Server, Oracle, and Elasticsearch, using native or agent-based drivers. It also includes message queues like Kafka, RocketMQ, RabbitMQ, Pulsar, and MQTT, as well as middleware consoles like Nacos, Consul, ZooKeeper, and etcd, enabling you to inspect infrastructure adjacent to databases in one tool.
AI SQL assistant and MCP server integration
In the query editor, select a table and describe what you want in natural language to generate SQL; the generated SQL undergoes a safety check before execution. It supports various AI backends, including Claude, OpenAI, and Ollama local models. Through the MCP (Model Context Protocol) server, AI coding agents like Claude Code, Cursor, and Windsurf can query databases using connections configured in DBX. The MCP server is deployed independently from the desktop app and allows setting three access modes: read-only, data read/write, and full access. A CLI package is also provided for use in the terminal, suitable for scripts and automation tasks.
Data grid, schema tools, and plugin extensions
The data grid, which applies virtual scrolling, handles large result sets and supports inline editing and export to CSV, JSON, Markdown, XLSX, and INSERT statements. The schema browser explores databases, schemas, tables, columns, indexes, foreign keys, and triggers, and visualizes table relationships with ER diagrams. It provides schema management tools such as schema diff comparison, execution plan visualization, and field lineage analysis. Redis and MongoDB have dedicated browsers that allow using system-specific features like key pattern search, document CRUD, and TTL editing. You can extend additional features like S3, Kubernetes, and LDAP as signed sandbox plugins from the plugin store.
Security settings per deployment environment and license
Connection information, AI settings, and tunnel credentials are stored encrypted, and desktop builds use platform-specific credential stores (macOS Keychain, Windows Credential Manager, Linux Secret Service). When deploying with Docker, the /app/data volume must be persisted, and in production environments, keys must be managed via DBX_SECRET_KEY_FILE or DBX_SECRET_KEY. For Windows portable builds, the entire data directory must be backed up, as credentials can be decrypted if the directory containing the keys is leaked. If credentials were stored in plaintext in a previous version, an encryption migration must be performed via the data security upgrade wizard on first launch. It is distributed under the Apache-2.0 license, and the development environment requires Node.js 18 or higher, pnpm, and Rust 1.88 or higher.
