TUI SQLite browser and query runner

A terminal UI app that lets developers open any SQLite file, browse tables and schemas, and run arbitrary SQL queries — all without leaving the command line or installing a GUI.

The idea

A single-binary CLI tool that opens a SQLite database file and presents a full terminal UI: a sidebar listing tables, a schema pane, a SQL editor, and a scrollable results grid. You run sqlite-tui myapp.db and immediately have a lightweight DBeaver-equivalent without leaving your terminal session. No Electron, no browser tab, no credentials to configure.

Why build this

SQLite is everywhere — mobile apps, embedded configs, local caches, test fixtures, Datasette exports, Obsidian vaults, browser profiles, n8n's default storage — but the tooling story for quickly poking around a file from the command line is poor. sqlite3 REPL works but has no visual structure. GUI tools like DB Browser for SQLite are fine on a desktop but unavailable over SSH, inside Docker, or on a headless VPS.

The gap is the headless/remote developer who just wants to run a few queries against a file without installing anything heavy. TUI frameworks have matured enough (Textual, Bubble Tea) that building a polished terminal app is now a weekend project, not a month of ncurses pain.

Stack sketch

Language: Python 3.11+
TUI framework: Textual — widget-based, CSS-driven layout, works over SSH
Database driver: sqlite3 from the standard library — zero dependencies for the core
Query results rendering: Textual's built-in DataTable widget with virtual scrolling
Distribution: PyPI (pip install sqlite-tui) and a single-file zipapp for no-install use
Testing: pytest + Textual's Pilot test driver for headless UI tests

Scope for v1

Open a SQLite file passed as a CLI argument; error clearly if the file is locked or corrupt.
Sidebar listing all tables and views, updated live if the file changes on disk.
Schema pane showing column names, types, and constraints for the selected table.
SQL editor panel with basic line editing (no syntax highlighting in v1).
Run query with a keybinding; display results in a paginated DataTable (default LIMIT 500, configurable).
Export the last query result to CSV with a single keystroke.
Read-only mode by default; a --write flag unlocks INSERT/UPDATE/DELETE.
Keyboard-only navigation throughout — no mouse dependency.

Out of scope for v1: multiple database connections, query history persistence, visual query builder, syntax highlighting, foreign key graph visualization.

Where it could go

The natural next step is multi-database support via SQLAlchemy — adding a --driver flag so the same UI works against PostgreSQL or MySQL over a connection string. That turns the tool from a SQLite-specific utility into a general headless database browser, which is a much larger audience. You could offer a hosted binary download and charge for the PostgreSQL/MySQL drivers as a one-time purchase.

A second expansion path is a "diff" subcommand: sqlite-tui diff a.db b.db that shows schema changes and row-level diffs between two files. This is immediately useful for anyone running migrations locally and wanting to verify what changed before committing, and it's a natural companion to the existing browse workflow.

Watch out for

Tables without an explicit ROWID or with millions of rows will make the DataTable unusable if you don't enforce a hard row cap — wire in a configurable LIMIT from day one and make sure the UI makes the cap visible so users don't assume they're seeing all rows.