Plans, translated ✓ ships
Execution plan findings in plain language, ranked by the cost they explain, each with its fix attached. Compare two plans side by side.
Every feature, sorted into three shelves: the table stakes, the premium stuff, and the genuinely new. Each one carries an honest mark — ✓ ships ◐ building ○ next — so nothing in progress reads as finished.
Every tool in this category has these. Ours just happen to be excellent.
SQL Server, PostgreSQL, MySQL, SQLite, DuckDB, ClickHouse, Oracle, libSQL/Turso, and Redis.
Each tab keeps its own connection; reopen the app and everything is exactly where you left it.
Whole document, selection, or current statement — and cancel actually stops the server-side query.
Keywords, tables, columns, aliases, and stored-procedure arguments as you type.
Output is re-lexed and verified token-for-token; anything off returns your SQL untouched.
Every run logged locally; named queries and templates one click away.
Find & replace, folding, multi-cursor, column select, split view, proper IME input.
CSV, JSON, Markdown, HTML, XML, SQL — and real .xlsx workbooks, not CSVs in a trench coat.
Drill-down tree, object definitions, script as CREATE / ALTER / DROP.
OS keyring storage, fail-closed TLS, Entra ID sign-in, one-file import/export.
Every command one fuzzy search away, rebound to your muscle memory.
One GPU-rendered app for macOS, Windows, and Linux.
The premium feel — details you’d start missing on day one anywhere else.
Execution plan findings in plain language, ranked by the cost they explain, each with its fix attached. Compare two plans side by side.
Right-click a row for a form built from the table’s own metadata — foreign-key dropdowns, CHECK constraints validated before the save leaves the app.
Compare two result sets row by row — adds, removes, and unchanged, with mismatched schemas called out. Name a snapshot and diff against it later.
Chart any result set, color-scale a column, add data bars — and export the chart as SVG.
JSON and XML open as collapsible, searchable trees; binary opens as a hex dump.
Activity monitor, backup/restore, schema compare, Query Store, Extended Events, template explorer, guided table/index/FK designers, procedure mode.
Group servers by client → project → environment, give each its own colors and icons, watch connection health, and wander the relationship graph.
Entra sign-in, subscription-grouped server discovery, and a firewall rule offered the moment the server says no.
History, snippets, connections, settings — plain files on disk. Grep them, back them up, drop them in git.
The actual app, compiled to WebAssembly, running in a tab against a sample database — before you download anything.
The genuinely novel parts. The reason this page exists.
The arrangement of your results lives in the query itself, as ordinary SQL comments: which columns lead, how they’re formatted, when they light up, what they chart. Arrange a result in the UI and SQLly writes the comment — open the file on any machine and it comes back identical. Read the docs
A cross-tab without a wizard, a separate tool, or a single line of PIVOT SQL. Two comments describe the rows, columns, and aggregate; the result grid reshapes itself and the arrangement travels with the file. Read the docs
Add a column to the result with a formula, not another pass of SQL. ! names refer to the columns you already selected; place the new column anywhere in the grid. Read the docs
Type the date the way you say it. SQLly writes the real literal and keeps your words beside it as a comment — and one command refreshes every dated literal against today, so a saved query never goes stale. Read the docs
Completions that read your data, not just your schema: finish a WHERE comparison from the column’s real values, and on a foreign-key column, the referenced names instead of bare ids. Read the docs
Your live database is the foundation; newer local DDL files win. SQLly swaps them in for IntelliSense, so you write SQL against the schema that’s coming — before it lands. Read the docs
Sketch a table that isn’t valid yet and SQLly overlays its best guess on the real schema — tagged as inferred, with its confidence, never passed off as authoritative. Read the docs
Require Touch ID or Windows Hello to connect — or before anything that isn’t a SELECT. Rules ride the same hierarchy as everything else: set it on a client, inherit it everywhere. Read the docs
Results stream in and spill to disk; the grid pages from disk while memory stays flat. Fifty rows or fifty million, the app doesn’t blink. Read the docs
Reach databases your machine can’t see — end-to-end encrypted through a relay that routes your traffic but can never read it. From any device, no VPN ritual. How it works
Also in the spells drawer: SQLly Embedded — the whole workbench inside your own web app ✓ · every IntelliSense bug report becomes a permanent regression test ✓ · live DDL git blame ○ · foreign keys complete as names, everywhere ✓ · the magic-string decoder ○
SQLly has a great deal more to offer than fits on one page. Wander the docs for everything else, then give it a try — it’s free.