Home - VheissuLabs/tql GitHub Wiki

tql (pronounced TEE-kwuhl, like sequel) is a database client for the terminal, for MySQL, MariaDB, PostgreSQL, SQL Server, SQLite, DuckDB, libSQL and Turso, ClickHouse, Redis and Valkey, and the hosted databases that speak their protocol. It is written in Rust and ships as one file.

Why

The database is the part of the job you cannot see. Everything else is in the terminal already (the editor, the logs, the deploy) and then the schema is behind a window you have to go and find, in an application that costs money and knows nothing about the rest of your work.

tql is that window, in the terminal, and it is built around three things a GUI tends not to do:

• It shows you the query. The SQL pane always holds the statement that produced what is on screen. Click a header to sort and the order by appears. Filter a column and the where appears, with its value bound, never glued in. You learn the language by using the tool.
• It is the same engine for you and for an agent. The tql query commands, the MCP server and the interface all go through the same code, so what your agent can see is what you can see. The agent's side only reads, and every statement either of you runs lands in one shared history. An agent can also put a statement in your SQL editor for you to read and run.
• It makes production look like production. A connection carries a tag and the tag carries a color, so the screen tells you where you are before you press d. Read-only connections refuse every write. Edits and deletions are pending until you type :w.

It is one binary, no configuration required to start, and it opens a SQLite file as happily as a Postgres server behind an SSH tunnel.

Install it

brew install vheissulabs/tap/tql            # macOS and Linux, Homebrew
curl -fsSL https://raw.githubusercontent.com/VheissuLabs/tql/main/install.sh | sh   # a script

There are .deb, .rpm and PKGBUILD packages, and standalone binaries, on the latest release. See Installation for all of them, including how to check a download.

Then run tql on its own to pick a connection, or point it straight at a database:

tql                                        # the connection list
tql ./database.sqlite                      # a SQLite file
tql 'mysql://user:[email protected]:3306/shop'

Databases

Database
MySQL and MariaDB	built in
PostgreSQL	built in
SQLite	built in, any .sqlite, .sqlite3 or .db file
SQL Server	built in, nothing else to install
DuckDB	built in, any .duckdb file, schemas and all
libSQL and Turso	built in, a libsql:// URL and its auth token (Connecting)
ClickHouse	built in, browse and query (Connecting)
Redis and Valkey	built in, a key browser of their own (Redis and Valkey)

A hosted database that speaks the MySQL or PostgreSQL protocol connects the same way, with a connection string or the connection form, over TLS and, if it needs one, an SSH tunnel:

• MySQL protocol: Laravel Cloud, Amazon RDS and Aurora, Google Cloud SQL, Azure Database, DigitalOcean, PlanetScale, TiDB, SingleStore
• PostgreSQL protocol: Neon, Supabase, Amazon RDS and Aurora, Google Cloud SQL, Azure Database, DigitalOcean, Crunchy Bridge, Timescale, CockroachDB

The wiki

Page
Installation	every way to install and update tql
Connecting	connection strings, the connection form, tags, groups, SSH and TLS
Browsing data	the grid, sorting, filtering, following links, structure
Editing data	cells, row forms, adding, pending changes, deleting
Redis and Valkey	the key browser: namespaces, every type, TTLs, the command pane
The row inspector	a row's fields and its related records
The SQL editor	simple and vim styles, history, completion
Asking for SQL	ask in english, and which model answers
Exporting	rows out as re-importable SQL
Dump, load and sync	whole databases, with mydumper, pg_dump and the rest
Commands	every tql subcommand and flag
Keys	every key, and how to rebind them
Configuration	every setting in config.toml
For agents	the tql commands and the MCP server
Troubleshooting	when something is not where you expected it
Developing	where the code lives, and the drawing rules
Packaging and releasing	the binary, the packages, the tap and the AUR
Changelog	what changed in each release

A database to try it on

The sample database used in development is Chinook, which is not committed to the repository. Fetch it and open it:

curl -L -o chinook.db https://github.com/lerocha/chinook-database/raw/master/ChinookDatabase/DataSources/Chinook_Sqlite.sqlite
tql chinook.db