Skip to content

Get Started with Opteryx

Opteryx is a cloud-native SQL query engine that lets you query data across multiple systems without needing to centralize or duplicate it. Whether your data lives in Postgres, Parquet, JSONL, or somewhere else, Opteryx gives you a unified SQL interface — so you can focus on analytics, not integration.

It's not a replacement for databases like Postgres or MySQL. Instead, Opteryx acts as a federated query layer, ideal for analytics workflows where data is distributed across multiple sources and fast answers matter more than transactional support.

Why Opteryx?

  • Query across formats and systems - Combine Parquet, Postgres, DuckDB, JSONL, and more in a single query. No ETL required.
  • Designed for the cloud - Opteryx is stateless and fast to start, making it perfect for serverless platforms like Google Cloud Run. No persistent state, no storage to manage.
  • Cost-efficient by design - Built for consumption-based infrastructure. Start fast, scale down to zero, pay only when queries are running.
  • Python-native - Use directly from Python or Jupyter notebooks. It implements PEP 249 (DB API v2.0), so it integrates seamlessly with your data science stack.
  • Time-travel support - Query your data as it was at a specific point in time — ideal for replaying decisions or debugging past behavior.

What Can You Use It For?

  • Analytics over distributed data - Run joins across logs in JSONL, user tables in Postgres, and reports in Parquet — without moving anything.
  • Data warehousing without the warehouse - Skip the data duplication step. Query data where it already lives.
  • Serverless data applications - Use Opteryx in cloud functions or ephemeral containers — no need for warm state or persistent infrastructure.
  • Decision auditing and retrospectives - Re-run your logic against historical data to understand how decisions were made.
  • Python-based pipelines - Drop Opteryx into your existing Python workflows. It behaves like a standard SQL engine, with native table and result access.

When Not to Use Opteryx

Opteryx isn't a general-purpose OLTP database. It doesn't support inserts, updates, or deletes, and it's not designed to replace your existing databases.

If you need:
- A full SQL engine with ACID guarantees → Use Postgres or MySQL
- An embedded analytics database → Use DuckDB or SQLite

We love those tools too — you can even query them using Opteryx.

Opteryx is for when you need to query across those tools, or want to run ad hoc queries on data in files.

Why Opteryx Rather Than 'X'?

'X' is great, and for your use case it may be a better option. Opteryx isn't for all users and all use cases - and we're okay with that.

Want to query:
- Customers in Postgres
- Web logs in JSONL
- Sales reports in Excel

…all in one query?
Opteryx handles that, with SQL you already know.

Architecture Fit

Opteryx is stateless between queries. That means:
- No session management
- No background jobs
- No persistent configuration

Just fast, predictable, read-only querying. It caches what it can, but if the container goes away, nothing breaks. Perfect for serverless, batch, or ephemeral compute environments.

⸻

Want to run a query now? Check out the Quickstart Guide
Want to integrate it into Python? See the Python Integration Docs