Why DataZen?

DataZen is a desktop database client built with Tauri + Rust. It combines the everyday database tools developers expect with AI-assisted querying, visual analysis, automation, and a compile-time driver architecture.

• Lightweight — Tauri + Rust keeps the application small and responsive.
• AI-native — generate SQL, diagnose errors, understand execution plans, and work with database context through chat.
• Visual — turn query results into charts without exporting to another tool.
• Automatable — compose SQL and AI operations into reusable YAML workflows across databases.
• Extensible — database drivers are integrated at compile time through the DataZen Driver API.
• Local-first — credentials stay on your machine (AES-256-GCM; master key in the OS keychain or a local .key file for unsigned/dev builds). Theme packs can customize DataTable type colors via --dt-* CSS tokens.
• Open source — GPLv3, with an architecture designed for community drivers and contributions.

A database client built around real workflows

SQL and data exploration

Write and run SQL in a modern editor, inspect results, browse tables, and move between query results and visualizations without leaving DataZen.

AI-assisted database work

DataZen puts AI next to the database instead of making you copy schema and errors into another application.

Natural language → SQL

Describe what you need and DataZen uses the current database schema as context to generate executable SQL. Generated SQL can be executed immediately or inserted into the editor for further editing.

SQL error diagnosis

When a query fails, AI can combine the database error and schema context to explain the problem and propose corrected SQL.

EXPLAIN analysis

Visualize execution plans and use AI to identify bottlenecks, scan strategies, and optimization opportunities.

Database-aware AI Chat

The AI sidebar can work with the current connection's schema and turn SQL from the conversation into editor-ready code.

Supported AI integrations include OpenAI, Anthropic, DeepSeek, and compatible custom endpoints.

Turn query results into charts

You should not need to export data to Excel just to understand it. DataZen can infer useful chart configurations from query results and switch between table and chart views.

Supported visualizations include line, bar, pie, scatter, and area charts, with aggregation, grouping, and PNG/SVG export.

Automate database work with Workflows

DataZen Workflows describe reusable database operations in YAML. A workflow can combine queries, AI steps, conditions, and loops, with each step connected to the database it needs.

For example, one workflow can query orders from PostgreSQL, fetch logistics from MySQL, and let AI summarize the combined result.

Workflows can be started from the UI, the AI sidebar, MCP, or generated with AI.

MCP: connect DataZen to the AI tool ecosystem

DataZen works both as an MCP Server and an MCP Client.

MCP Server

Expose database operations, schema inspection, EXPLAIN, and workflows to external AI agents. DataZen also provides a headless stdio mode for automation and agent integrations.

MCP Client

Connect external MCP servers to DataZen AI Chat and bring additional tools and context into database conversations.

1. Open Settings → External MCP Servers and add a server (command, optional args/env).
2. Enable Expose to AI Chat so the assistant can call that server's tools.
3. Connect the server (or restart DataZen — enabled servers auto-reconnect on launch).
4. In AI Chat, ask naturally; when the model needs an external capability it calls tools named mcp/{serverId}/{toolName} alongside built-in database tools.

This makes DataZen useful not only as a GUI, but also as a database tool inside larger AI-assisted development workflows.

Extensible database drivers

DataZen separates the application from database-specific implementation through the DataZen Driver API.

                         DataZen
                            │
              ┌─────────────┴─────────────┐
              │       DataZen Core        │
              │  UI · Query · AI · MCP   │
              └─────────────┬─────────────┘
                            │
                    DataZen Driver API
                            │
          ┌─────────────────┼─────────────────┐
          │                 │                 │
       PostgreSQL         MySQL          External drivers
                                           │
                              ┌────────────┼────────────┐
                              │            │            │
                           MongoDB      ClickHouse    OLAP...

Drivers are compiled into DataZen rather than loaded through an unstable Rust dynamic-library ABI. This allows a driver to provide both Rust database functionality and frontend UI while still remaining in its own repository.

Independent driver development

A driver can be developed in an independent repository next to a local DataZen checkout:

workspace/
├── datazen/
└── datazen-driver-mydb/

During development, DataZen's driver registry can point to the local repository with source: "path". The DataZen application is then built with the selected driver, giving plugin developers a real host for both backend and frontend debugging.

The Driver API itself is maintained in packages/driver-api and published as the MIT-licensed datazen-driver-api crate. Independent drivers normally consume the published crate and do not need to clone DataZen just to obtain the API.

See the complete guides:

• Independent Plugin Development — English
• 独立插件开发指南 — 中文
• Driver API crate README
• Driver API dependency boundary
• datazen-driver-api on crates.io

Supported databases

DataZen ships with a small default set and can be built with additional drivers.

The exact driver set is controlled at build time, so a distribution does not have to ship every database engine.

Installation

Download the latest release from GitHub Releases.

DataZen is free and does not require an account.

macOS Gatekeeper: If the app is blocked as damaged or from an unidentified developer, clear quarantine: xattr -cr /Applications/DataZen.app, or right-click → Open once. See packaging.md for details and the notarization checklist.

Optional drivers (MongoDB, ClickHouse, DuckDB, SQL Server, …) are compile-time selections — see optional-drivers.md.

Build from source

Prerequisites

• Node.js >= 20
• pnpm >= 9
• Rust >= 1.77
• Tauri v2 system dependencies

pnpm install
pnpm tauri dev

Build only the drivers you need:

# Default driver set
pnpm tauri:build

# All supported path drivers
DATAZEN_DRIVERS=all pnpm tauri:build

# Custom driver set
DATAZEN_DRIVERS=postgres,mongodb pnpm tauri:build

Security and privacy

DataZen is designed around local database access:

• Database credentials are stored locally.
• AI requests are sent to the provider configured by the user.
• Database data is not uploaded to a DataZen cloud service.
• SSH connections can be established directly from the application.

Always review the privacy and security policies of the AI provider and endpoint you configure.

Documentation

• Project website · User Manual (EN) · 使用手册 (ZH)
• Feature guides · Architecture docs · Development & release docs
• Independent Plugin Development
• Chinese Plugin Development Guide
• Driver API crate
• Driver API dependency boundary
• datazen-driver-api on crates.io
• Workflow Guide
• Optional path drivers (MongoDB, ClickHouse, DuckDB, SQL Server)
• Packaging & release channels (macOS Gatekeeper, Linux packages)
• Auto-update
• Contributing

Contributing

DataZen welcomes bug reports, feature requests, database drivers, documentation improvements, and code contributions.

Please read CONTRIBUTING.md before opening a pull request. Driver work should generally be developed in an independent driver repository and integrated through the DataZen driver registry.

License

DataZen is licensed under the GNU General Public License v3.0. The datazen-driver-api crate under packages/driver-api is separately licensed under the MIT License. See LICENSE and packages/driver-api/LICENSE-MIT.