Documentation

Query your data,
where it lives.

Opteryx is a small, fast SQL engine that reads data in place — files on S3 or GCS, Iceberg tables, PostgreSQL, Firestore, or a dataframe in your process. Nothing to load first.

Start here · 5 min

Your first query, in less time than this paragraph.

Install with pip, point at a Parquet file, and pull rows back. No cluster, no schema upfront — just SQL.

Open quickstart

# pip install opteryx-coreSELECT name, mass FROM READ_PARQUET('planets.parquet') WHERE mass > 1.0 ORDER BY mass DESC LIMIT 3;

┌─────────┬────────┐│ name │ mass │├─────────┼────────┤│ Jupiter │ 1898.0 ││ Saturn │ 568.0 ││ Neptune │ 102.0 │└─────────┴────────┘

// reference

SQL Reference

Statements, expressions, functions, operators, and data types — the full Opteryx grammar.

SELECT JOINs Window functions Aggregates String fns Types →

// reference

API Reference

Embed Opteryx in Python, or call it over HTTP. Connection objects, cursors, and results.

opteryx.query Connections Cursors POST /v1/query Errors →

Stuck? Talk to a human.

The Opteryx community hangs out on Discord — maintainers included. Most questions get answered within the hour.

Join the Discord

Query your data,where it lives.

Your first query, in less time than this paragraph.

SQL Reference

API Reference

Stuck? Talk to a human.

Query your data,
where it lives.