DataZen vs SQLly: a 10 MB Tauri client against a deep SQL workbench
An honest look at DataZen, the free GPL database client built with Tauri and Rust: download sizes, engines, workflows, MCP write modes, and where it beats SQLly.
DataZen is a free, open-source database client built with Tauri, Rust and React. Its pitch is size: the default Windows installer is about 10 MB, and inside that it fits a schema-aware SQL editor, a Redis workbench, YAML workflows, dashboards with alerts, and an MCP server. It overlaps with SQLly in a lot of places and makes very different trade-offs in others, so this post sets the two side by side.
We build SQLly, so take the bias into account. We read DataZen’s README, user manual, changelog and website, and checked the files in its 0.2.2 release on GitHub. We did not install or benchmark it. For several kinds of user it is the better choice, and we say which ones below.
Small on purpose
DataZen’s website says “under 15 MB”, and the release files back that up for the default build. In decimal megabytes:
| Download | SQLly | DataZen 0.2.2 |
|---|---|---|
| macOS, Apple Silicon | 65.6 MB dmg | 12.5 MB dmg (22.5 MB for the All build) |
| Windows x64 | 60.9 MB zip | 10.2 MB installer, 9.9 MB portable zip (17.1 MB All) |
| Linux x64 | 76.2 MB tarball | 13.8 MB deb or rpm, 91.1 MB AppImage |
Two choices make it that small. The interface is a React app running in the webview the operating system already ships, so DataZen doesn’t carry its own renderer. And the default build has only four drivers: PostgreSQL, MySQL/MariaDB, SQLite and Redis. A second “All” build adds the rest.
SQLly goes the other way. Its interface is Rust drawn on the GPU through Metal, Direct3D or Vulkan, with no browser engine, and every engine, DuckDB included, is in the one download. That costs about 50 MB more. If download size is what you care about most, DataZen wins and it isn’t close.
Databases
| Engine | SQLly | DataZen |
|---|---|---|
| PostgreSQL, MySQL, MariaDB, SQLite, Redis | Yes | Yes, in every build |
| SQL Server | Yes, plus Azure SQL (Entra sign-in in progress) | All build only |
| ClickHouse, DuckDB, Turso | Yes | All build only |
| Oracle Database | Yes | Not in its driver list |
| MongoDB, Elasticsearch, InfluxDB, HBase, vector databases | No | All build only |
| Valkey, KeyDB, DragonflyDB, ElastiCache, MemoryDB | Documented | Not documented |
DataZen covers more kinds of database. If you need a document store, a search engine or a time-series database next to your SQL, its All build has drivers for them and SQLly has none. SQLly covers more SQL servers in one build, and per engine it goes further: plans, activity monitors, dashboards and compare tools on each one. The engine list spells out which tools work where.
One detail from DataZen’s changelog is worth knowing before you migrate data with it: only PostgreSQL-to-PostgreSQL, MySQL-to-MySQL and PostgreSQL-to-MySQL pairs are verified end to end for Data Sync and Data Transfer. Other pairs are labeled experimental in the app.
Writing SQL
DataZen’s editor is better than most free clients. Completions come from live metadata, hover shows column types and comments, function calls get signature help, a linter flags errors as you type, and Alt+Enter offers quick fixes. It has named parameters, snippets, transaction controls, a visual query builder, and natural-language SQL through the AI provider you pick.
SQLly’s IntelliSense goes a step past
metadata. It tracks aliases and scope, writes a JOIN ON clause from real foreign keys,
pulls a column’s actual values into a WHERE clause, understands plain-language
dates, and flags a GROUP BY mistake while you type. The editor handles a 0.5 MB script
at about 5 ms per keystroke, with templates, multi-cursor editing, split views, folding and AI
ghost text.
A licensing note: DataZen’s README says a build from source compiles its open-source editor, while the published installers include the complete editor. The release files include a “SQL Editor Pro” extension that is pulled from a separate repository, which we couldn’t find in public. If you plan to build DataZen yourself, expect the editor to differ from the download.
Results
DataZen streams rows as they arrive, edits cells inline on tables with a primary key, turns a result into a line, bar, pie, scatter or area chart, and exports CSV, TSV, JSON, Markdown, XLSX or SQL.
SQLly’s grid spills to disk past 50,000 rows or 64 MiB, so fifty million rows use about as much memory as fifty. On top of it are pivots, charts, conditional formatting, computed columns, result diffs and Excel export with live formulas, and a one-line comment saves the layout with the query. We couldn’t find pivots or result diffs in DataZen’s documentation.
Automation is DataZen’s strongest card
This is the part of DataZen with no real SQLly equivalent. Workflows are YAML files that chain queries, driver commands, AI steps, conditions and loops, and each step can target a different connection, so one workflow can read from PostgreSQL, read from MySQL, and summarize both. You can start them from the UI, from MCP, or on a schedule. Ops Dashboards are saved canvases of charts bound to SQL that refresh on an interval and can raise a desktop notification or call a webhook when a value crosses a threshold.
SQLly has scheduled tasks and a Server Dashboard, but no workflow language and no alert webhooks. If you want your database client to run chores while you’re away, DataZen does more.
AI and MCP
Both apps work with cloud AI providers or a local Ollama model, and neither turns AI on until you configure it. DataZen lists OpenAI, Anthropic, DeepSeek and any OpenAI-compatible endpoint. DataZen uses AI for natural-language SQL, error diagnosis and plan analysis. SQLly also runs on-device models through MLX.
Both ship an MCP server, and their permission models differ:
- DataZen lets agents write, if you allow it. Its server exposes schema, query, EXPLAIN and workflow tools. Each tool can be switched off, connections can be allowlisted, and the permission mode is read-only, safe write or high-risk write. It also runs headless over stdio for agent pipelines, and DataZen can act as an MCP client, pulling other servers’ tools into its own chat.
- SQLly never lets an agent write.
sqlly-cli mcplets an agent list connections, read schemas and run oneSELECTat a time, inside a transaction the database itself keeps read-only, which is then rolled back. If the agent wants to change data, it sends a proposal; SQLly shows you the exact SQL and you run it or decline it. Production connections are refused outright.
If you want an agent that runs migrations or fixes rows, DataZen can do it. If you want agents to look and never touch, SQLly has no mode to switch on by mistake.
Safety
DataZen has the basics a person needs: read-only connections that reject writes in the driver,
and a Safe Mode that blocks UPDATE or DELETE without WHERE,
plus TRUNCATE and DROP. Saved passwords are encrypted with AES-256-GCM
and a master key in the system keychain.
SQLly’s guardrails go further: SELECT-only connections that also block file-reading and shell-escape tricks, automatic rollback, tenant checks, an optional AI review before a data change, and production painted red. You set them on a client, project or environment, and every connection under it inherits them.
Admin tools
DataZen has an ER diagram, schema diff with a DDL plan, data sync, cross-engine transfer, scheduled backups, a server status view with a process list you can kill from, and a Redis workbench with a live MONITOR. That’s a solid set for a free app.
SQLly adds the DBA side: Activity Monitor with blocking chains and waits, Query Insights over Query Store, pg_stat_statements and Performance Schema, Extended Events and SQL Agent for SQL Server, an index analyzer, a test data generator, disk usage, risk-graded schema compare with rollback scripts, and query plans on nine engines (still in progress). Anything destructive opens as SQL in a tab for you to read first.
Platforms, polish and maturity
DataZen runs on Intel Macs as well as Apple Silicon, back to macOS 10.15, and its interface comes in ten languages. SQLly needs Apple Silicon and speaks English and Spanish. DataZen’s Windows and Linux builds are x64 only; SQLly ships amd64 and arm64.
DataZen’s download page says its Mac app isn’t notarized yet and that Windows SmartScreen
may warn, so first launch means a right-click or an xattr command. SQLly’s Mac
build is signed and notarized. DataZen is also young: its first GitHub release was in August 2026,
the latest is 0.2.2, and one maintainer wrote nearly all of its commits, at a fast pace. That pace
is a good sign, but expect things to move.
What they cost
DataZen is free, needs no account, and is GPL-3.0-or-later, with an MIT-licensed Driver API and a linking exception for plugins, themes and drivers. SQLly has a free tier for hobbyists (two saved servers, and not for organizations over 100 people), a one-time $100 SQLly 1.x license, and a $200 Lifetime license; see pricing. Lifetime licenses carry a source-escrow promise if SQLly is ever discontinued, but SQLly is not open source. If open source or free for the whole company is a requirement, DataZen meets it and SQLly does not.
Which one to pick
Pick DataZen if:
- You want the smallest download you can get, or you need open source, or free at any company size.
- You need MongoDB, Elasticsearch, InfluxDB or HBase next to your SQL databases.
- You want YAML workflows, scheduled automation, or dashboards that alert a webhook.
- You’re on an Intel Mac, or want the interface in one of its ten languages.
- You want an AI agent that can write, or you want to extend the app with drivers and themes.
Pick SQLly if:
- You work in SQL Server, Azure SQL, Oracle, ClickHouse or DuckDB and want them in one build.
- You write a lot of SQL and want completions that know your data, not only your schema.
- You pull large results and want a grid that pivots, diffs and stays flat on memory.
- You read plans, watch blocking, or run Query Store and SQL Agent.
- You want agents to read and never write, and a signed, notarized Mac app.
The fair test is your own database. Neither app needs an account to start. The side-by-side table is at SQLly vs DataZen, and SQLly’s download is free.