Command-line access

The qore CLI gives you scriptable, headless access to the connections you already saved in the QoreDB desktop app. It runs the same engine, reads the same vault, and enforces the same safety gates — so a query that's safe in the app is safe from the terminal, and there's no separate credential setup.

This makes QoreDB usable in places the desktop UI can't reach: CI pipelines, cron jobs, remote servers over SSH, and quick one-off checks without leaving your shell.

How it works

• Shared credentials. Connections and their passwords come from the same OS keyring and config directory as the desktop app (the vault, project default). If the desktop can connect, so can the CLI — nothing to re-enter.
• Same safety engine. Queries go through the existing safety layer (production-dangerous blocking, the anti-loop rate limiter, governance limits). A connection keeps its saved read-only flag.
• Machine-friendly output. Results are printed as JSON on stdout. Errors go to stderr with a non-zero exit code, so scripts can branch on success.

Commands

qore connections                                      # list saved connections
qore query <connection_id> "<sql>"                    # run a query, print rows as JSON
qore tables <connection_id> <database> [--schema s]   # list tables/collections
qore describe <connection_id> <database> <table> [--schema s]

A typical session starts by listing connections to get an id, then querying:

qore connections
qore query prod-pg "SELECT count(*) FROM orders WHERE status = 'paid'"
qore tables prod-pg analytics --schema public
qore describe prod-pg analytics orders --schema public

Pointing at a different config

By default the CLI reads the desktop app's config directory. Set QOREDB_CONFIG_DIR to point it elsewhere — useful on a CI runner where you've staged a dedicated config:

QOREDB_CONFIG_DIR=/etc/qoredb qore query prod-pg "SELECT 1"

Building it

The CLI ships as a separate binary you build from the QoreDB workspace:

cargo build -p qore-cli --release   # produces target/release/qore

Where to go next

• MCP server to expose the same connections to AI agents
• Vault & profiles for how credentials are stored
• Best practices for safe production access