Overview

RED helps you inspect schemas, browse large tables, run SQL queries, and export data without the complexity of heavyweight IDE-style database tools.

Roughly Enough Data. Show enough to make decisions quickly.

RED is rendered natively on the GPU - no Electron, no browser runtime, no web stack. It connects to SQLite, PostgreSQL, MySQL / MariaDB, ClickHouse, Redis / Valkey, and MongoDB.

Three shells, one app

RED is not a SQL tool with bolt-ons. Each engine family gets a workspace shaped like the data it holds, sharing the same chrome, keymap, settings, and AI panel:

• SQL shell - schema tree, query tabs, result grid. SQLite, PostgreSQL, MySQL / MariaDB, ClickHouse.
• Redis shell - a key browser with Browse, Console, Pub/Sub, Monitor, Analysis and Keyspace panels. See Redis.
• MongoDB shell - a database → collection tree with Documents, Query, Schema and Indexes panels. See MongoDB.

What's here

• Schema explorer - namespaces, tables, views, materialized views, functions, procedures, triggers, sequences and user-defined types, loaded lazily per group.
• SQL editor - query tabs with autocompletion, FK-aware JOIN suggestions, inline diagnostics, EXPLAIN, formatting, and saved snippets.
• Large-result browsing - a windowed grid over multi-million-row results with keyset pagination. Nothing is materialized whole.
• Result tooling - find, filter to a WHERE clause, resize / hide / reorder / freeze columns, pin rows, and inspect a cell or row.
• Inline editing - staged, reviewable batch changes with per-statement confirmation, plus manual transactions.
• AI assistant - an opt-in sidebar grounded in the connected database, over your own API key or a Claude subscription, with a tiered permission model.
• Safety by environment - mark a connection Local / Dev / Staging / Prod and the write guards follow. See Safety and guards.
• Panes and tabs - any number of panes as columns, rows, or nested splits, with draggable tabs.
• Multiple connections - hold several at once and switch keyboard-first; reach a database over an SSH jump host or a SOCKS5 / HTTP proxy.
• Data export - CSV, JSON, SQL INSERT, Excel .xlsx, and HTML, all streamed row by row.
• Server insight - a Server panel per engine with overview metrics, live sessions, and a guarded kill.
• Command-line mode - run queries, execute scripts, seed dev/staging databases, and serve MCP tools, all headless.

Next steps

• Installation - download a release or build from source.
• Connecting to a database - every engine, TLS, SSH, proxies, imports.
• The SQL editor - query tabs, running, panes, and history.
• The result grid - browsing, filtering, and editing rows.
• Redis - the key browser and its panels.
• MongoDB - the document shell.
• AI assistant - grounded chat over your schema.
• Safety and guards - environments, confirmations, transactions.
• Exporting data - get results out of RED.
• Keyboard reference - drive RED almost entirely from the keyboard.
• Command-line mode - script queries, seed databases, serve MCP.