rainfrog and SQLly: a terminal database tool and a desktop SQL workbench
rainfrog is an MIT-licensed Rust TUI with vim keys that runs over SSH; SQLly is a GPU-drawn desktop workbench with SQL Server, deep IntelliSense and DBA tools.
rainfrog describes itself in five words: “a database tool for the terminal”. It’s a Rust program built on Ratatui, MIT-licensed, with about 5,400 stars on GitHub and a release roughly every few weeks (v0.4.6 shipped on 24 September 2026). You browse schemas and tables, write SQL with vim-style keys, keep a history and a list of favorites, and copy or export what comes back. SQLly is also written in Rust, so people ask how the two compare.
We build SQLly, so take the bias into account. We read rainfrog’s README, skimmed its source for how it handles writes, and checked the sizes of its release assets. We did not benchmark it or use it against a production database. For a fair number of people it’s the better tool, and we say who below.
Terminal vs desktop
This is the real difference, and everything else follows from it.
- rainfrog lives in a terminal. You can SSH into the box next to the database and run it there, keep it in a tmux pane, or start it from a script with a
--urlor aDATABASE_URLfrom a.envfile. Connections can also live in a TOML config file, and passwords can go into the platform keychain. - SQLly is a desktop app. The window is drawn on the GPU through Metal, Direct3D or Vulkan. That buys a grid, charts and dialogs a terminal can’t draw, and it costs you the SSH case entirely: SQLly needs a graphical session.
sqlly-cliruns in a terminal for command-line use and as an MCP server, but there is no full-screen terminal client.
Size follows the same split. rainfrog’s v0.4.6 tarballs are 18.9 MB for Apple Silicon, 16.3 MB for Windows x64, 18.1 MB for Linux x64, and 8.0 MB for the Linux musl build, which leaves DuckDB out. SQLly’s downloads are 65.6 MB (macOS dmg), 60.9 MB (Windows zip) and 76.2 MB (Linux tarball). rainfrog also installs from cargo, Homebrew, Arch’s official repositories, nix, conda, pixi or Docker, and runs on Intel Macs, 32-bit Linux and Android under Termux. SQLly needs Apple Silicon on the Mac.
Databases
| Engine | SQLly | rainfrog |
|---|---|---|
| PostgreSQL | Yes | Yes (tier 1, most tested) |
| MySQL, SQLite | Yes | Yes (tier 2) |
| DuckDB | Yes, bundled | Yes (tier 2), not in the Android, i686 or musl builds |
| Oracle | Yes | Tier 3, needs Oracle Instant Client |
| Amazon Redshift | Not documented | Tier 2, through its PostgreSQL driver |
| SQL Server, Azure SQL | Yes (Azure SQL Entra sign-in in progress) | No |
| MariaDB, ClickHouse, libSQL/Turso, Redis | Yes | No |
rainfrog’s README is upfront about its tiers, which we like. If you work in Postgres, it’s on solid ground. If you work in SQL Server, it isn’t an option. SQLly’s engine list covers 30 engines and says which tools work on each one.
Writing SQL
rainfrog’s editor is vim-flavoured: normal, insert and visual modes, motions like
w, b and gg, yank and paste, undo and redo. The README notes that it
doesn’t behave exactly like vim. Completions cover SQL keywords, schema names, words in the buffer
and file paths. Alt+e sends the query to your $EDITOR, so you can write it in
real vim or Neovim and come back. Each run takes one statement; F7 skips the parser for
anything it can’t handle.
SQLly puts much of its work here. IntelliSense
tracks aliases and scope, writes a JOIN ON clause from real foreign keys, pulls a
column’s actual values into a WHERE clause, and flags a GROUP BY
mistake while you type. It runs whole scripts, including GO batches, and stays at about
5 ms per keystroke on a 0.5 MB script. There’s a Vim mode in the editor and an Open in External
Editor command, but the rest of the app is mouse and shortcuts, not modal keys.
Results
rainfrog shows results in a scrollable table you move through with hjkl, select a cell
or row, and copy. Results export to CSV in your Downloads folder; the README warns that export blocks
the app, so a big one freezes it until it finishes. You can’t edit cells.
SQLly’s grid streams rows to disk past 50,000 rows or 64 MiB, so 50 million rows use about as much memory as 50. On top of it are pivots, charts, conditional formatting, computed columns, result diffs, Excel export and a row editor.
Safety
rainfrog has a sensible default we’d point anyone to. An UPDATE or
DELETE runs inside a transaction, and rainfrog asks whether to commit, telling you how many
rows it touched. DROP, TRUNCATE and ALTER ask before they run.
Even so, its README says not to use it with write access on a production database.
SQLly’s guardrails go further: SELECT-only connections that
block file-reading and shell-escape tricks, a hard stop on UPDATE or DELETE
with no WHERE, automatic rollback, tenant checks, an optional AI review before a data
change, and production painted red. You set them once on a client, project or environment and every
connection under it inherits them.
Everything else
rainfrog stops at querying, and that’s a choice, not a gap it’s hiding. SQLly has the DBA set under one Tools menu: Activity Monitor, Query Insights, query plans (still in progress), schema and data compare, cross-engine data transfer, an ER graph, SQL Agent, and backup and restore as scripts you read before running. It also has AI help with local or opt-in cloud models and an MCP server that only ever reads.
What they cost
rainfrog is free and MIT-licensed, with no paid tier. SQLly has a free tier for hobbyists (two saved servers), a one-time $100 SQLly 1.x license, and a $200 Lifetime license; see pricing. If open source or “free for the whole company” is a requirement, rainfrog meets it and SQLly doesn’t.
Which one to pick
Pick rainfrog if:
- You query databases from SSH sessions or tmux.
- You want vim keys everywhere and your own
$EDITORone keystroke away. - You want an 8-20 MB tool that installs from your package manager.
- You need open source, or free at any company size.
- You mostly work in PostgreSQL and want a quick look at tables.
Pick SQLly if:
- You work in SQL Server, MariaDB, ClickHouse, libSQL or Redis.
- You write long SQL and want completions that know your data.
- You analyse results: pivots, charts, diffs, Excel.
- You want guardrails stronger than a confirm prompt, especially near production.
- You read plans, compare schemas, or watch server activity.
Plenty of people will want both: rainfrog on the server, SQLly on the laptop. The side-by-side table is at SQLly vs rainfrog, and SQLly’s download is free.