// sqlly vs harlequin

SQLly vs Harlequin

Harlequin is “the SQL IDE for your terminal”: a free, MIT-licensed Python app with a data catalog, a tabbed editor and a results table, all inside the terminal you already have open. It connects through adapter plug-ins, starting with DuckDB. SQLly is a desktop app with a GPU-drawn window. The real question here is terminal vs desktop, and each one wins different jobs. Here’s the honest breakdown.

Last reviewed: November 2026

The short version

SQLly

A desktop SQL workbench

One Rust codebase drawn on the GPU on macOS, Windows, and Linux. Thirty engines with native drivers, IntelliSense that reads your data, a grid built for analysis, layered guardrails, and a full set of DBA tools. Agents get read-only access; anything that writes comes back to you as a proposal.

Harlequin

A SQL IDE that lives in your terminal

A Textual app installed from PyPI that runs over SSH, in tmux, or anywhere Python 3.10+ runs. DuckDB is built in, and adapter packages add Postgres, MySQL, Snowflake, BigQuery, Trino and more. It comes with hsql, a headless companion CLI for scripts and AI agents. Version 2.17.0 shipped on 26 October 2026.

Feature by feature

SQLly’s column uses the same honest status badges as the features page and the docs. For Harlequin we list what its own docs, README, and changelog state.

Feature comparison of SQLly and Harlequin
Capability SQLly Harlequin
Runs in a terminal or over SSH NoA desktop app that needs a display. sqlly-cli covers command-line queries and MCP, with no terminal UI. YesThe whole app is a terminal UI, mouse included.
Runs on macOS, Windows & Linux YesmacOS is signed and notarized; Windows and Linux are a supported beta, on amd64 and arm64. YesAnywhere Python 3.10-3.14 runs, Intel Macs included.
Install DownloadA 65.6 MB dmg on macOS, a 60.9 MB zip on Windows, a 76.2 MB tarball on Linux. Mac App Store edition too. PyPIuv tool install harlequin is the recommended route; pip, pipx and a community Homebrew formula also work.
Free to use Free tierTwo saved servers, one cloud account, and a reminder to pay about every five hours. Pricing YesFree at any company size, supported by GitHub Sponsors.
Open source NoLifetime licenses carry a source-escrow promise if SQLly is ever discontinued. YesMIT on GitHub.
DuckDB & SQLite YesDuckDB is bundled. Per-engine coverage YesBoth adapters ship with Harlequin; DuckDB is the default, with MotherDuck and extensions.
PostgreSQL, MySQL & MariaDB YesPlus documented wire-compatibles such as CockroachDB, Neon, Supabase, and Aurora. YesSeparate adapter packages from Harlequin’s maintainer, installed as extras.
SQL Server & Oracle YesNative drivers, nothing extra to install. Azure SQL too (Entra sign-in in progress). PartialThrough the ODBC adapter, which needs a driver manager and the vendor’s ODBC driver.
ClickHouse, libSQL/Turso & Redis YesRedis covers Valkey, KeyDB, DragonflyDB, ElastiCache and MemoryDB. Redis PartialA community chDB adapter for embedded ClickHouse. We found no Redis or libSQL adapter.
Snowflake, BigQuery, Trino, Databricks, Redshift PlannedSnowflake, BigQuery and Trino are planned; none of these connect today. YesSnowflake and Redshift from the maintainer; BigQuery, Trino and Databricks from the community.
Cassandra, NebulaGraph, Exasol, H2 No YesCommunity adapters, plus ADBC for any Arrow-driver database.
SSH tunnels YesSSH hops, SOCKS and HTTP proxies, and Kubernetes. Engine docs YesUses your installed ssh client, so your whole ~/.ssh/config applies.
Schema browser YesServer explorer with DDL, designers, and per-object tools. Server explorer YesA Data Catalog tree with type labels and an Interactions menu (Describe, Show DDL), plus local-file and S3 trees.
Autocomplete YesAlias- and scope-aware, JOIN ON from foreign keys, real column values in WHERE. IntelliSense PartialKeyword and catalog completions with exact and fuzzy matches.
SQL formatter YesFormatter YesF4 formats the buffer with sqlfmt.
Query history YesSearchable, with pinned snippets. Query History YesF8 opens a searchable history that also shows what hsql ran.
Large results YesStreams and spills to disk past 50,000 rows or 64 MiB; flat memory at 50 million rows. Streaming PartialAn Arrow-backed table that shows up to 100,000 rows by default; the cap is configurable.
Edit table data YesA row editor generated from the table’s own metadata. Row editor -We couldn’t find grid editing in its docs; changes are made with SQL.
Pivot, charts & result diff YesPivot, charts, conditional formatting, computed columns, and diffs on any result. Pivot No
Export PartialCSV, TSV, JSON, Markdown, SQL, Excel with live formulas, and code generators. No Parquet from the grid. Export PartialCSV, Parquet, JSON, ORC, and Feather; copy as TSV. No Excel.
Themes & keybindings YesTwelve themes in light and dark; rebindable keys. Themes · Keybindings YesAny Textual theme; a default vscode keymap, with custom keymaps from TOML or plug-ins.
Transactions YesManual-commit toggle, commit and rollback keys, and auto-rollback guardrails. PartialCommit and rollback buttons on adapters that offer more than one transaction mode.
Read-only connections YesSELECT-only connections that block file-read and shell-escape tricks. Select-only Yes--read-only asks the database to refuse writes, and refuses to start if the adapter can’t enforce it.
Other guardrails in the editor YesA hard stop on UPDATE/DELETE without WHERE, tenant checks, prod painted red. Safety -Beyond read-only, a row limit and a timeout, we found none documented.
Execution plans, activity monitor, schema compare YesActivity Monitor, schema and data compare, Query Insights; plan viewer on nine engines is in progress. Tools NoHarlequin is an editor, catalog and results viewer; it doesn’t try to be a DBA tool.
AI agents YesAn MCP server whose SELECT runs read-only and is always rolled back; writes come back as proposals. How reads stay read-only Yeshsql, a headless CLI with a 500-row default limit, --read-only, --timeout, and a Claude Code skill. We found no MCP server.
AI in the editor YesLocal models, your own endpoint, or opt-in cloud providers. On-device AI -We found no AI chat or completion inside Harlequin itself.

A dash means we couldn’t confirm it from Harlequin’s own materials - check its docs at harlequin.sh for the current state.

Where SQLly is different

Harlequin keeps you in the terminal and reaches the warehouses. SQLly gives you a full window and goes deep on each engine it supports.

A grid that does analysis

Pivot, charts, conditional formatting, computed columns, result diffs, and Excel export with live formulas and a pivot sheet. A one-line SQL comment can save the layout so it travels with the query.

SQLly directives →

Results past 100,000 rows

SQLly streams rows and spills to disk past 50,000 rows or 64 MiB, so fifty million rows use about the same memory as fifty. You scroll the whole result instead of a capped window.

Spill-to-disk results →

Autocomplete that knows your data

SQLly can pull a column’s real distinct values into a WHERE clause, write a join’s ON clause from actual foreign keys, and underline a GROUP BY mistake as you type.

IntelliSense →

Native drivers, no ODBC setup

SQL Server, Oracle, ClickHouse, and Redis connect with drivers built into the app. There’s no driver manager or vendor ODBC package to install first.

Databases →

Guardrails beyond read-only

A hard stop on UPDATE/DELETE with no WHERE, auto-wrapped transactions, tenant checks, an optional AI review before a data change, and prod painted red - inherited down your client, project, and environment tree.

Safety & guardrails →

DBA tools in the same app

Activity monitor with blocking chains, Query Store and pg_stat_statements, SQL Agent, schema and data compare, index analyzer, test data, and backup and restore as scripts you read before they run.

Tools →

When Harlequin is the better choice

Harlequin is probably the better fit if:

  • You work on remote boxes. SSH into a server, run harlequin, and you have a catalog, an editor and a results table right there. SQLly needs a desktop, so you’d tunnel to the database instead.
  • You live in tmux and the keyboard. Harlequin sits in a pane next to your shell and editor, and its keymaps are plain TOML.
  • You query a warehouse. Snowflake, BigQuery, Redshift, Trino and Databricks all have adapters. SQLly doesn’t connect to any of them today.
  • Your data is Parquet and DuckDB. DuckDB is the default engine, the file tree shows local files and S3 buckets, and results export to Parquet, ORC and Feather. SQLly bundles DuckDB, but its export dialog has no Parquet option.
  • You want a small, scriptable, free tool. The wheel is about 356 KB on top of its Python dependencies, it’s MIT, it’s free at any company size, and hsql gives scripts and agents the same adapters and config. SQLly’s source is closed, and its free tier isn’t available inside organizations with more than 100 employees.

Plenty of people will use both: Harlequin on the server and in the terminal, a desktop workbench for analysis and admin work.

Questions

Is Harlequin free?

Yes. Harlequin is open source under the MIT license, free at any company size, and supported by GitHub Sponsors. You install it from PyPI with uv, pip or pipx, or from a community-maintained Homebrew formula. SQLly is free for hobbyists, with a one-time SQLly 1.x license and a Lifetime license for everyone else.

Which databases do SQLly and Harlequin support?

Harlequin ships adapters for DuckDB and SQLite. Its maintainer also publishes Postgres, MySQL/MariaDB, Redshift, Snowflake and ODBC adapters (ODBC covers SQL Server, Oracle and others), and the community maintains BigQuery, Trino, Databricks, ADBC, chDB, Cassandra and more. SQLly has native drivers for 30 engines, including SQL Server, Oracle, ClickHouse and Redis, with no ODBC driver to install, but it does not connect to Snowflake, BigQuery, Trino, Databricks or Cassandra today.

Does SQLly run in a terminal like Harlequin?

No. SQLly is a desktop app drawn on the GPU, so it needs a display. Its sqlly-cli command runs queries and serves MCP from the command line, but there is no terminal interface. If you work over SSH on a server or inside tmux, Harlequin is the better fit; it runs anywhere Python 3.10 or later does.

// try it yourself

Point both at the same database

The quickest way to choose is to run your own queries in each. SQLly is free on macOS, Windows, and Linux, with no account required.