// a database tool that actually likes you back

SQLly

pronounced SKWEE-lee  Β·  /ˈskwiːli/

The friendly SQL RDBMS tool that makes querying, editing, and babysitting your databases genuinely pleasant. Built by a nerdy pig with glasses who loves a well-normalized table.

SQLly, the nerdy pig mascot with glasses
-- oink
SELECT * FROM
users WHERE cool;
CREATE TABLE joy (
  id SERIAL
);
// try it now, no download

Run SQLly right here in your browser

The real workbench compiles to WebAssembly and runs entirely in your tab - nothing to install, no account, no server round-trips. Query two live sample databases and poke at every button before you download a thing.

sqlly.app/app
SELECT name, count(*) AS tracks
FROM "Artist" JOIN "Album" USING (artist_id)
GROUP BY name ORDER BY tracks DESC;
Launch the browser preview β†’

Real SQLite, in the tab

The Northwind sample runs on SQLite compiled to WebAssembly - actual queries, actual results, no emulator.

Real Postgres, via PGlite

The Chinook sample is a full Postgres running in your browser through PGlite - the same engine, no cloud.

Stays on your machine

Everything runs locally in the tab. This labeled preview can't reach your servers - the desktop app is the real thing.

// not ai slop

Human designed. AI assisted. Human verified.

Yes, LLMs write a lot of the code here. No, that doesn’t mean we ship whatever falls out of the prompt. Every feature runs the same gauntlet:

  1. 1

    Spec first, code second

    Every feature starts as a hand-written implementation spec: what it does, where the edge cases hide, and exactly how we’ll know it works. No prompt-and-pray.

  2. 2

    Then stress-test the spec

    Before any code is written, an adversarial review tears the spec apart - digging up extra edge cases, potential security issues, and second-order effects worth keeping in mind.

  3. 3

    Tests define “done”

    Tests are often written by hand before any implementation exists. The LLM’s job is to make them pass - it doesn’t get to grade its own homework.

  4. 4

    A committee of models writes the code

    Implementation is split across Opus, Kimi K3, GLM 5.2, GPT 5.6 Sol, and MiniMax M3.

  5. 5

    Then the models argue

    Adversarial reviews pit the models against each other, and once the code is in, a separate verification pass nitpicks and double-checks everything. None of them pull punches.

  6. 6

    A human has the last word

    Every change gets a final human review before it ships. The committee advises; the human decides.

The gauntlet is helped along by dont-get-fooled-again, an opencode plugin we built ourselves to aid exactly this process - it haggles over the proposed solution, then verifies the implementation. And plenty of features skip the committee entirely - some code is just more fun to write by hand.

// features

Everything a database deserves

Powerful enough for the DBA who never sleeps, friendly enough for the dev who’s still Googling “how to join.”

⚑

Query Editor

A lightning-fast editor with smart autocomplete, syntax highlighting, and results that don’t make you wait for coffee.

πŸ“

Edit data, not just SELECT it

Right-click any result row for a generated edit form - foreign-key dropdowns, constraint checks, insert, duplicate, delete. No hand-written UPDATEs for a one-cell fix.

πŸ—„οΈ

SQL Server & Azure SQL, first-class

Deep T-SQL support with real Entra ID auth - device code or browser, no service-principal secrets. More engines are on the roadmap.

πŸ“ˆ

Performance insights

Spot slow queries before production notices. Execution plans with full operator-level detail - the graphical plan view is in progress.

πŸ”

Export every which way

CSV, TSV, JSON, Markdown, HTML, XML, SQL INSERTs - or a real .xlsx that opens as a formatted Excel table. No 2 a.m. reformatting scripts.

πŸ”’

Sensible Security

Credentials stored locally and encrypted. No telemetry phoning home - the pig minds its own business.

Explore all features β†’
// why sqlly

Why a pig would choose SQLly

There are a hundred database tools. Here’s why this is the one that sticks.

1

Fast, and we mean it

Native, lightweight, no Electron bloat. It opens before you finish saying “SKWEE-lee.”

2

Friendly, not fiddly

A UI designed by humans πŸ§‘β€πŸ’» (and one bespectacled pig 🐷) cough cough and AI πŸ€–. Intellisense that knows what you want.

3

Honest pricing

Free forever, fairly priced to go nag-free. No surprise enterprise upsells lurking behind a friendly button.

4

Built by people who care

Made by developers who got tired of clunky clients. Every feature exists because somebody actually wanted it.

3
desktop platforms, one codebase
500ms
cold launch time
0
trackers or ads
∞
happy little tables
// totally real praise*

People are squealing about it

*Testimonials are, uh, enthusiastic reenactments. But the love is real - read the real reviews.

“I used to dread opening my database client. With SQLly I actually look forward to it. My therapist says that’s growth.”
Norm A. LyzedBackend Engineer, Hash & Sons
“The row editor alone replaced three tools and one spreadsheet I was too embarrassed to mention. Worth every cent.”
Rollo VerbackDBA, Codd Enterprises
“Fast, clean, and the autocomplete actually finishes my thought before I do. The pig vibes are a bonus.”
Portia TionData Engineer, ACID Analytics
“Finally, a database tool built by someone who has clearly suffered through database tools. Every button is right where my hand already was. It’s almost suspicious.”
Cash K. CadedPlatform Lead, B-Tree Mobility
“I’m not a DBA - I just know enough SQL to be dangerous. SQLly does way more than hand me a table to copy into Excel. It actually helps me understand what I’m looking at.”
Orland DerbyOperations Manager, Shard Logistics
“My subqueries have subqueries. SQLly formats the whole mess so I can still read it tomorrow. Ten out of ten, would nest again.”
Sue B. QueryAnalytics Lead, Outer Join Collective
SQLly mascot
// about us

A pig, a dream, and a lot of tables

SQLly started the way most good tools do: with frustration. We were tired of bloated database clients that felt designed by a committee that had never written a JOIN. So we built our own - and we put a bespectacled pig on the box because, honestly, why not.

We believe database tools should be fast, friendly, and a little bit fun. Serious enough for production, approachable enough for a Monday morning.

βœ“

Simple by default. Sensible settings out of the box, power features where you expect them.

βœ“

Respectful of your machine. Tiny download, low memory, no background nonsense.

βœ“

Respectful of you. No dark patterns, no upsell walls, no pig-nonsense.