Skip to content
 
 

Repository files navigation

RED

RED

Roughly Enough Data. A fast, native database explorer built in Rust.

Release Downloads CI Stars License Platforms Rust

RED querying a MySQL database, with the schema tree, result grid, and AI assistant sidebar

RED is for inspecting schemas, browsing large tables, running SQL, and exporting data. The name is the goal: show roughly enough data to make a decision quickly, without the weight of a full IDE-style database tool.

It's built on GPUI and renders on the GPU, so there's no Electron or embedded browser.

Status: MVP. The core workflows are functional, but the UI and APIs are still moving. Expect rough edges and breaking changes before the first stable release.

Databases

  • PostgreSQL, MySQL / MariaDB, SQLite: browsing, querying, and transactional in-grid editing
  • ClickHouse: browsing, querying, bulk insert, and in-grid editing through background mutations you can watch and cancel
  • Redis: key browser, value lenses and decoders, console, keyspace notifications, slowlog and monitor
  • MongoDB: document workspace with inferred schema, aggregation, explain, and document editing

Features

Browsing and querying

  • Schema explorer covering tables, views, materialized views, functions, procedures, triggers, sequences and types
  • SQL editor with autocompletion, foreign-key-aware JOIN suggestions, inline diagnostics, and per-statement run
  • Large tables stream through a windowed, keyset-paginated result grid instead of being materialized whole
  • Find in results and in the editor; filter a result to a WHERE clause without rewriting the query
  • Cell and row detail inspector, per-column statistics, split view of two tabs
  • Watch mode: re-run a read-only query every few seconds, with changed cells flashing
  • Query history and saved queries

Editing and moving data

  • Inline editing with staged, reviewable batch changes
  • Export to CSV, JSON, SQL INSERT statements, or a self-contained HTML report; import from CSV, JSONL, or a JSON array
  • Copy a table to another connection, or migrate a set of tables into a new database
  • Diff data between two tables, and compare schemas with reconciling DDL generated into a query tab

Understanding a server

  • ER diagram built from foreign keys, with pan, zoom and fit
  • Connection health report: sizes, unused indexes, missing keys, and other per-engine checks
  • Server panel listing live sessions, with lock waits and their blockers marked and a guarded cancel or terminate
  • Show DDL for any object in the schema tree

Working against production

  • Read-only connections
  • Environment markers (Local, Dev, Staging, Prod) that scale how much RED asks before a write
  • Destructive statements are graded before they run, and the confirmation says how many rows they will touch

Everything else

  • AI assistant sidebar with grounded chat over your schema, via the Claude API or a Claude subscription
  • SSH tunneling through a jump host, SOCKS5 / HTTP proxies, and TLS
  • Import saved connections from DBeaver and DBGate
  • Multiple connections with a quick connection switcher and pinned favourites
  • Themes (One Dark, GitHub Dark, Ayu Dark, Ayu Light, High Contrast) and a fully customizable keymap
  • Headless CLI (red query, exec, copy, migrate, test, connections) and an MCP server (red mcp)

Default shortcuts are listed in docs/keyboard.md, and every release is recorded in CHANGELOG.md.

Install

Prebuilt, signed binaries are on the latest release:

  • macOS: download the .dmg (signed and notarized).
  • Linux: download the .AppImage, chmod +x it, and run.
  • Windows: download the .exe, or the .zip if you prefer to unpack it.

Or build from source; see Development.

On first launch RED seeds a small, read-only Sample database so you can explore the schema browser, run queries, and try the result grid immediately, with no database setup required.

Privacy

RED has no telemetry and makes no network calls of its own. Connection credentials are stored in your operating system's keychain, never in plaintext. The AI assistant is opt-in and only talks to the provider you configure.

Architecture

A GPUI main thread renders the interface while a Tokio backend service owns database sessions and query execution. The UI and backend communicate through a Command/Event channel bridge, so the interface never blocks on the database.

Workspace crates:

  • red: desktop application
  • red-core: shared domain types
  • red-driver: database driver abstractions and implementations
  • red-service: backend runtime and query lifecycle
  • red-config: saved connections and OS-keychain access, shared by the app and the CLI
  • red-ai / red-acp: AI assistant providers (direct API and agent-client-protocol)

UI components and theming come from Flint, a shared component library built on GPUI.

Development

cargo run -p red
cargo test
cargo clippy --workspace --all-targets -- -D warnings
cargo fmt --all

On macOS, building requires a full Xcode installation for Metal shader compilation. If xcode-select points to the Command Line Tools, set:

DEVELOPER_DIR=/Applications/Xcode.app/Contents/Developer

See CONTRIBUTING.md for development setup and contribution guidelines.

License

GPL-3.0-or-later.

RED links against GPUI, whose dependency tree includes GPL-licensed crates. See NOTICE for details.

About

A fast, native database explorer, featuring an AI assistant that understands your schema, drafts queries for approval, and turns results into charts.

Resources

Code of conduct

Contributing

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages