RustGrid
Guide
Benchmark
Download
  • English
  • 简体中文
GitHub
Guide
Benchmark
Download
  • English
  • 简体中文
GitHub
RustGrid

RustGrid

A lightweight, native database GUI in Rust. Navicat layout, zero learning curve.

DownloadQuick start

One native binary

Built with Rust + GPUI. No Electron, no runtime, ~10 MB download.

Familiar by design

Connection tree, data grid, row form and dialogs follow Navicat.

Nine engines + ODBC

MySQL, MariaDB, PostgreSQL, SQL Server, Oracle, SQLite, ClickHouse, Redis, MongoDB and ODBC.

SQL editor

Highlighting, completion, Format SQL, Explain, saved and group queries.

Backup & sync

Single-file .rgbak backups and a table-level diff you can preview before writing.

Made to extend

Drivers, connection pages and table designers are declarative specs, not UI forks.

Size

release 0.1.15

The complete desktop build ships as one native executable — no runtime, no Electron.

9.8MBWindows x64 download
Windows ARM648.2 MB
Windows x649.8 MB
Linux ARM6411.3 MB
Linux x6413.1 MB
macOS Apple Silicon13.1 MB
macOS Intel14.5 MB
Installed9.8 MBone executable
Cold start745 msuntil the window appears
Idle memory88 MBworking set

Archive (zip / tar.gz) sizes from the GitHub release. Unpacked, the Windows build is a single RustGrid.exe.

Performance

grid operations, lower is better

Every number is a grid operation through RustGrid’s own driver stack, on the same six-column table and the same machine.

RustGrid paints the page first and counts the total in a separate task, so first-page time excludes COUNT(*).

10,000,000 rows
First page · 1,000 rows
rows only; COUNT is separate
MySQL 2.81 ms
PostgreSQL 4.50 ms
ClickHouse 6.70 ms
MongoDB 5.81 ms
Oracle 8.31 ms
SQLite fastest2.55 ms
Page 101 · offset 100,000
deep pagination
MySQL 16.9 ms
PostgreSQL 31.6 ms
ClickHouse fastest9.44 ms
MongoDB 25.0 ms
Oracle 31.0 ms
SQLite 13.3 ms
Page 1,001 · offset 1,000,000
deep pagination
MySQL 1,716 ms
PostgreSQL 276 ms
ClickHouse fastest15.0 ms
MongoDB 195 ms
Oracle 221 ms
SQLite 112 ms
Primary-key lookup
average of 30 lookups
MySQL 2.16 ms
PostgreSQL 2.49 ms
ClickHouse 4.59 ms
MongoDB 0.49 ms
Oracle 4.99 ms
SQLite fastest0.23 ms
Filter + sort · 100 rows
status = paid, newest first
MySQL 2.23 ms
PostgreSQL 3.45 ms
ClickHouse 9.82 ms
MongoDB 1,819 ms
Oracle 3.91 ms
SQLite fastest1.94 ms
COUNT(*)
full table
MySQL 2,606 ms
PostgreSQL 148 ms
ClickHouse fastest3.20 ms
MongoDB 1,129 ms
Oracle 126 ms
SQLite 192 ms
All operations
EngineFirst pageOffset 100kOffset 1MPoint lookupFilter + sortCOUNT(*)
MySQL 8.4.132.8116.91,7162.162.232,606
PostgreSQL 17.114.5031.62762.493.45148
ClickHouse 26.10.4.2076.709.4415.04.599.823.20
MongoDB 8.3.135.8125.01950.491,8191,129
Oracle 23.26.3.0.08.3131.02214.993.91126
SQLite 3.54.12.5513.31120.231.94192
Machine Windows 11 · WSL2 · 12 GB RAM · 6 vCPU
Table id, user_id, status, amount, created_at, note
Data 1,000,000 / 10,000,000 rows · index (status, created_at)
Engines MySQL · PostgreSQL · ClickHouse · MongoDB · Oracle · SQLite
Method 1 warm-up + 3 runs, median · tools/rgbench

Databases ran one at a time in Docker on one machine, each with a warm-up; numbers are indicative, not production absolutes.

SQL Server, MariaDB and Redis are not part of this run.


Read the guide: Connections · Tables · Functions · Queries · Backup · Sync · Settings · Keyboard shortcuts

Apache-2.0 · Copyright © 2026 yisier