omni-sql

Features

Analyze files and database results locally

Received a CSV or Parquet export to investigate? Import it and query it with SQL in the same IDE, using embedded DuckDB, without loading it into your source database or writing a separate script. You can also import a query result from any supported connection and join it with the file locally.

1. Run a source query and choose Analyze locally: a full result or a sample.
2. Add a CSV or Parquet file (with a delimiter preview) and join it with SQL.
3. Inspect the result and export the full local query output.

Local datasets persist across restarts, and S3 objects (CSV, Parquet, Delta, Iceberg, DuckLake) can be queried and imported too. A sample stays a sample: exports never restore rows left out at import time. Try the demo with fictional data or read the local analysis guide.

SQL editing and exploration

• Metadata-aware completion with PK/FK badges and types, including columns projected by CTEs; preview a single CTE's rows with Ctrl+Alt+Enter.
• Dialect quick fixes, formatting (Ctrl+Alt+L) and /catalog SQL templates.
• Execution plans for PostgreSQL, MySQL/MariaDB, SQL Server and Oracle.
• Filter, sort, page and export results; edit rows when a primary key makes the update safe.
• MongoDB: native Extended JSON queries, a document view, and read-only SQL with /mongo conversion. See MongoDB usage.

Connections

Organize connections in folders, tag them with your own environments (Development, Staging, Production or custom ones, each with a color), and import them from an omni-sql export or from DBeaver. Passwords stay in the OS keyring and are never exported.

AI assistants through MCP

A local MCP server lets compatible clients (Codex, Claude Desktop, ChatGPT Desktop) read the active SQL tab, inspect schemas and plans, propose edits, and run SQL. Edits and executions always need approval in the desktop, and each connection can be set to Off, Read-only (default) or Read and write.

See the MCP guide for setup and the security model.

Database support

Database	Connection	Metadata autocomplete	Query execution
PostgreSQL	✅	✅	✅
MySQL	✅	✅	✅
MariaDB	✅	✅	✅
SQL Server	✅	✅	✅
Oracle	✅	✅	✅
MongoDB	✅	✅ Sampled fields	Native Extended JSON + read-only SQL
Generic JDBC	🧪 Experimental	Driver-dependent	Driver-dependent
ODBC	🧪 Experimental	Driver-dependent	Driver-dependent

Generic JDBC needs a user-supplied driver JAR; ODBC needs a separately installed 64-bit driver. Neither offers plans, definitions or row editing. Details and limitations: Database support.

Install

Download your package from the latest release.

Platform	Package
Windows 10/11 x64	.exe installer or x64-portable.zip
Debian/Ubuntu amd64	.deb
macOS 15+ Apple Silicon / Intel	aarch64.dmg / x64.dmg

On Windows you can also install with WinGet:

winget install cccadet.omni-sql

Node.js, Java and the ODBC driver manager are bundled; native connections need no database client. Release assets include SHA256SUMS. Linux ARM, AppImage and RPM are not packaged yet.

macOS builds are ad-hoc signed and not notarized: on first launch use System Settings → Privacy & Security → Open Anyway. The installation guide covers this, the Windows portable archive, where data is stored, and updates.

omni-sql is early-stage software. Try it with development data before using it against important environments, and please report unexpected behavior.

Quick start

• Files only: choose Import file, select a CSV or Parquet file, and query it in the Local DuckDB tab.
• A database: create a connection, choose Test connection, connect, then browse objects or open a SQL tab. Run the selection or current statement and use Analyze locally to bring the result into DuckDB.

Stuck? See Troubleshooting.

When to use omni-sql

omni-sql focuses on SQL editing plus local analysis of database results and files. Mature tools such as DBeaver and DataGrip cover broader administration; you can use omni-sql alongside them.

Choose omni-sql for…	Consider a broader tool for…
Local DuckDB joins across database results and files	Deep vendor-specific administration
One editor across five relational databases and MongoDB	Data modeling and migration tools
CTE- and metadata-aware completion	A large plugin ecosystem or enterprise support
A local, MIT-licensed app with no account	Team features and a long-established product

Documentation

Read the documentation site: installation, getting started, the SQL editor, connections, local analysis, MongoDB, MCP, shortcuts, troubleshooting and contributor guides. The same pages are in docs/.

Built with Tauri, React, Fluent UI, Monaco Editor, TypeScript, Rust, and Kotlin. The interface is available in English, Portuguese and Simplified Chinese.

Contributing

If omni-sql is useful to you, star the repository. Report a bug, share a database compatibility result, or tell us about your first run. Pull requests are welcome; read CONTRIBUTING.md first.

License

MIT