Multi-driver
One client for relational, document, and graph databases — from Postgres to Snowflake — each behind a common browse-and-query surface, so muscle memory carries across engines.
Chitra is a native, multi-driver database client for the terminal: browse schemas, run queries in DBeaver-style tabs that share one connection, visualize your schema as an ER diagram, and keep every credential in a Touch ID-guarded vault — installed standalone or through the Swavan CLI.
Chitra installs two ways: a one-line standalone installer (no swavan CLI
needed), or through the swavan CLI alongside your other plugins. Either
way it self-updates.
curl -fsSL https://raw.githubusercontent.com/swavan/plugins/main/install-chitra.sh | sh
irm https://raw.githubusercontent.com/swavan/plugins/main/install-chitra.ps1 | iex
swavan plugin install chitra
swavan launch chitra
chitra --self-update # standalone update
swavan plugin update chitra
swavan plugin uninstall chitra
The standalone installer resolves the latest version from the public catalog,
SHA-256-verifies the signed download, and only replaces the binary when a newer build
exists — so re-running it (or chitra --self-update) is a safe no-op when
you're already current, and never downgrades a newer local build.
A single terminal window: a collapsible schema navigator, tabbed query editors that share one connection, result grids, and an in-terminal ER diagram.
One client for relational, document, and graph databases — from Postgres to Snowflake — each behind a common browse-and-query surface, so muscle memory carries across engines.
Open multiple query tabs that all share a single connection. Write, run, and keep several working queries side by side without juggling sessions.
The left panel groups objects into collapsible schema folders under an Entities header, so large databases stay navigable. Drill in to load table detail on demand.
Render a focused table's relationships or the whole schema as a box-and-connector diagram drawn right in the terminal, with foreign-key edges — plus a Mermaid export for copy/save.
Query results render in a scrollable grid that preserves each column's real type, so numbers, timestamps, JSON, and text all display faithfully.
An in-TUI Settings ▸ Keybindings screen lets you search, remap, and live-rebind every key, then save — so the whole surface bends to your habits.
These are the defaults — every binding is remappable from Settings ▸ Keybindings (F1), with search and live rebinding.
Chitra is keyboard-first: the command surface, schema navigator, query tabs, and result grid are all reachable without the mouse. If a default clashes with your terminal or multiplexer, remap it in F1 and save.
Connection credentials live in a local encrypted vault, unlocked with Touch ID where available — and Chitra can reconnect the moment it starts.
Saved connections are held in an encrypted vault that unlocks with a single biometric prompt on supported machines — credentials never sit in plaintext config, and startup never silently reads the keychain.
Turn on auto-connect (Settings ▸ General) and Chitra reopens your chosen connection on launch, so you land in the schema browser instead of a connection picker.
Every connection carries its own user-configurable timeout, and credentials are scoped to the vault — so sharing a machine or scripting a launch never leaks secrets into your shell history or dotfiles.
Chitra is a TUI first, but a few flags handle version, updates, and help straight from the shell.
chitra # start the TUI
swavan launch chitra # via the CLI
chitra --version
chitra --self-update # standalone
swavan plugin update chitra
chitra --help
Installed standalone, chitra keeps itself current with
--self-update; installed as a plugin, it updates with the rest of your set
via swavan plugin update.